Ballroom H Saturday Mar. 07 - SCaLE 23x

This talk discusses open-source software licensing, particularly the implications of switching from open-source to business source licenses by various companies. The speaker examines the reactions within the developer community, highlighting concerns about transparency and community involvement when such transitions occur. Community forks, like OpenSearch and Valkyrie, are noted as responses to these shifts, emphasizing community attempts to maintain open-source principles. The speaker also analyzes how open-source foundations operate financially, detailing the sustenance model based on membership fees and the role of member companies in providing both funding and developer resources. Lastly, he addresses misconceptions about the influx of contributors following a project’s transition to a foundation, explaining that introducing a project to a foundation does not guarantee immediate increases in developer engagement.

add me on the mic. It's green. >> Yep. Hello. Am I miked or no? Okay. All right. Well, I mean I I project anyway, so I think I'll be just fine. Uh you really think it's that low? I can talk as loud as you want, but I'm worried about the recording. >> Yeah. Yeah. Usually that's what you're never supposed to do with microphones. Uh just put them

too close to your mouth. Anyway, if you can hear me, okay, I'm not worried about it. The recording is picking up fine, I'm sure. Um so, uh you know, I just uh for those of you that were in the room and apparently on the live stream as well, um if you saw on the live stream anything you weren't supposed to see, apparently I was displaying lots of

my screen because they told me it wasn't going through. But uh so if uh so I will need to ask them to edit the live stream edit as soon as possible to cut that out. Um but follow along my slides because I they my machine to make HDMI work which doesn't scale organizers that was the uh the AV for the venue who didn't know what they were

talking about but then I accidentally shut down my web server because I also had a root window open to that and so now so you can't follow along my slides online but they will be up on my on a URL uh at some point in the future. Um, so the original plan for this talk was I was going to explain what happened at trial in the Vizio

case because our trial was scheduled if you came to my talk last year for September of 2025. It was then rescheduled for January 2026 and uh an hour before the pre-trial hearing. So the pre-trial hearing is on a Thursday. The trial starts on a Monday. Is all there. We're all down here. Not here but in Orange County. and uh the judge clerk called the lawyers an hour

before our hearing and said she had taken the trial off the calendar. Uh in California there are various rules about who gets trials first uh there is a thing called a five-year rule uh which is if cases pending for five years it must go to trial before ones that aren't. And then within that jury trials are prioritized over bench trials. We were having a bench trial in

our case. So, uh, there was a five-year jury trial, which is way ahead of us because we're we're just not yet at the five-year mark. Uh, and we are also a bench trial. So, her trial ran over. She canceled it. We now are, we know, well, I'll tell you more about that in a minute. But what I want to start with is ask everybody to consider masking.

I haven't had a mask on mainly because I've been pushing around with this AV stuff, but most of the day you'll see me wearing a mask. I think it's a good I mean, that it's supposed to have two things, >> you saying do Yeah, it's pretty un It's coming out of the speakers. Yeah, I can't. This thing will not get closer to me. So, >> so I

can try it this way, too. >> That's the way I had it before. It's it it works on this ear, but not that ear, but okay. It it for some reason it I had it both ways on this side and it didn't work. Oh well. Okay. Well, anyway, uh I would like to encourage all of you to to mask trickle in the trade show floor and so

forth. Uh that's what I'm doing. Um we also have a booth at the Sol Freedom Conservancy. It is sadly a proprietary firmware, but it tells you the air quality. The air is actually very good in here because of the high ceiling. So it's not as uh risky. It's kind of a proxy reading. Uh but uh at big events like this uh it's very important. I I went

to FASD in 2025. I wore a mask the entire time because FOSSDM you think this is crowded. FOSM is the most crowded event in open source and free software in the world every year. And um I got COVID and I couldn't work for a month. So I mean CO didn't kill me. I recovered fine but free software lost a month of my work. So, and if you

think about all the people that go to Fosom that work on free software and how many of them got CO and how bad free and open source software was hurt by people not wearing masks at FASDM, this is why I still encourage best events. So, before I get deep into the Vizio case, especially because I don't have I was hoping to give you uh how the trial

went uh I am going to update on something that was in the news just this week. And because it's California specific, I thought it was really important to spend some time. So, I'm sorry you're getting not as advertised for a little bit of this content, but I think it's more valuable because a lot of it's going to be recycled from last year because not not much has

changed in the video case since then. So, uh a lot of people got concerned about this age verification law uh last week. Uh uh the reason why is some BSD sub project no one has ever heard of announced that they were not going to distribute their software anymore in California which caused a number of poorly written tech journalist articles that were searched um saying the sky was

falling in California with regard to open source and free software operating systems. Um the law was 2025. See it keeps moving. I can't keep it in the place that you want it. So, it's the best I can do. Um, the law was passed October 2025 and signed into law I I really don't like using a handheld mic. So, I will do this way. Um, it's move again.

So, that it'll be good for a while and then it'll change again. Um, but this is relatively loud. Yeah. I really hate the handle. I'm gonna try this one more time. But if that's if that's the only option, it's okay. I'll just use the handle. I don't So, okay. Well, my talk quality is going to go down because I don't do very well with a handheld mic.

So, sorry about that. Um, so the uh this this law was passed in October 2025. It was signed into law and it doesn't go into effect until January 1st, 2027. Now, the date January 1st, 2026 is mentioned in some of the details about the law. And my theory is is that basically people were confused and thought it went into effect back in January. One key thing to

note is it's a business regulation, not a criminal statute. So it's not like if you don't comply with this law, you're going to go to prison. Um you can be fined by the state of California. And as I said, it doesn't go into effect for a while now. So we have a lot of time to deal with this thing. And I really do think that project that

announced the sky was falling and that they had to withdraw their project from California entirely, I really really think they were um uh they were just overreacting to this law. So I don't think you should panic about it at all. If you've heard about it, don't worry about it. We've done some preliminary analysis at the Software Freedom Conservancy. Um and uh we're planning to publish a more

comprehensive document. Obviously, this just became a big news item this week. We didn't see it as a top priority given the January time frame. Um we've got a whole Vizio trial to do before then. So it wasn't our first priority obviously, but we've now made it a higher priority because the world seems to be freaking out about it. Um but one of the things I in the

research I've done this week I discovered is there have been no regulations issue yet. And if you're familiar at all with how state governments tend to operate, I live in Oregon and it's very similar there. Um the the everything's in the details, right? So, normally what happens is a statute is passed and then there's a the regulatory body of the state gets together and writes much more

detailed regulations of how they're actually going to enforce that particular legislation. And I did comprehensive research. There is no pending discussion about any regulations for this law. There's none planned and there's none been issued. So, we really should wait until we get some regulation, especially and if there isn't some, I'm really not that worried because California also has this really neat thing that I just learned about

this week, which is really I usually hold questions at the end if it's absolutely urgent. I'm happy to take it, but I'm already just I'm already flustered enough. Um, so I'd really appreciate not getting opted. Okay. Well, is it okay if my whole question is is that okay with you? I mean, I I'm already extremely flustered because of all the AV problems. I'd really just like to

finish my talk and then I'm happy to take any questions you have and I will be in the booth for the entire rest of the weekend. So, California actually forbids what's called underground regulation. And this is a case, this is a situation where um someone uh someone in the state government uh writes something down, writes a rule down and starts following it or otherwise directs people to

follow certain rules. um you actually have the uh ability to petition as citizens of California uh if you believe underground regulations have occurred and and get that red addressed uh by the government. So So if this law goes into effect and if it starts being enforced right away and there have been no actual official regulations issued uh we'll have this opportunity to f to to figure out

if there's underground regulations happening. Um the other thing to note about the law is that it is not is not DRM. It's not restricted boot. It's not saying you have to give a verified boot chain that verified age that cannot be modified. Says nothing like that. It does use the word secure secure method. And it's clear from context. They mean you have to transmit the age verification

data over HTTPS or something similar. So it really just means encrypt it when you send it to the other side. Um and I've looked into like how we actually implement this. Um, I'm pretty sure you could implement this on Debian and comply if you just anytime somebody adds a user, which means we probably have to put an installer because the installer let you add a user. Um,

you you select an age range, which your age range is, and then you just patch Firefox and Chromium to have a JavaScript API that can answer that, and Debbian would be in compliance if it had that with the law. Um and also there's nothing in the law that makes the operating system vendor as it's defined in the law um responsible for verify like getting an ID like

in some states people are having to like scan in IDs to and upload them to get to websites. That's not what this law says. The user can lie to you and it is not your fault and you have no responsibility to figure out the user is lying to you. Ironically, the person making the app, right? So, Google in the app store, they might have they might have

to worry about that, but you don't. Because if they have a because and here's why. There's this goofy little thing in the law that says if by other means you've done age verification, then you are required to use that data instead of what you got from the operating system. So, if Google got the kid's ID when they set up the phone and Google knows it's a 15-y

old using the device and Google doesn't correctly correct you so that the 15-year-old says, "Oh, I'm 50." And then you say, "Oh, user said they were 50." And Google knows the user is really 15 because it's an Android that's like some sort of Android app store using Google Google Play apps, Google Play app store. Um, Google gets in trouble for that, not you, not Debbian or or

Android in that case. So, um I I I think I'm surprised Misilla is not paying attention to this because uh their their plug-in archive would for be example where they're kind of like an app store and they might be covered by the app store provisions, but nobody who makes operating systems is going to be covered by the app store part of the provisions. Um the other thing

that happened is when signing it, I I don't know why governors do this. I've seen a lot of them do it. They they don't like a bill for political reasons because it was to save the children um from horrible whatever online. They sign it and then they issue one of these publishing signing statements which is this is a link but obviously my website's down right now so

you can't click on the link but it's if you search for it you'll find it quickly. Um and he is obviously very concerned about how the how the bill is drafted. So for all we know, the legislature could come back into session and just basically overturn the law this year uh with something else or just repeal it all together. All that could poss within realm of possibility.

So that's the content on that. Uh we we will be publishing at SFC a comprehensive analysis that is sort of our our quick take analysis uh that given a week of wall clock time to do. Um and I'm happy to take questions on it at the end of the talk. Um, but I want to move on to the Vizio case, which is really now the Walmart case.

Uh, because during our lawsuit, Walmart acquired Vizio for $2 billion. Um, they are now using that money to pay very expensive lawyers to stop us from getting the source code. Um, Walmart is uh, uh, Vizio V. Vizio is a wholly owned subsidiary now. Walmart. So, they are it's just like Red Hat and IBM. There is no real Red Hat anymore. It's just IBM and they have a

division called Red Hat. Walmart now has a division called Vizio. And it's very fascinating to think about that that there's, you know, this is the large then we're now fighting the largest retailer in the world who has decided they don't want to give us source code. I want to go briefly over uh I'll go relatively quickly through this because I know some of you at my talk

last year, which we talked about this last year as well. Um, we initially got a report that Vizio was failing to comply with the GPL in 2012 and we investigated that and then it looked like they weren't. We moved on to other matters. 2018 we got another uh report and my colleague Denver Gener who's here. He's also speaking later today on a different track. I hope you

go to his talk as well. Um, he decided to open what we call open the matter. Um, so we basically decided to send them a cease and desist letter, ask them to redress and fix the problems uh of not giving any source code. In fact, they had no offer for source code whatsoever. So they have to make an offer and then have to honor that offer. And

Denver spent most of 2019 uh sorry that's supposed to be reviews six candidates um from in from January through December 2019. They go back and forth with these what we call the complete corresponding source code candidate. What that is is we tell the company, get us the source code and they give us something that they say is the source code. It's never actually the source code. It

is some source code for some things that happen to be in the firmware which usually don't build. And then we go back to them and give them a list of problems with that candidate. Then they send us another candidate and it has all the same problems and then more. And then we go back and forth and and we were on six when we got there. The most

by the way I've ever had in a matter is 28. um before they got into compliance. Um so finally in December that year 2019 uh note the date by the way that's important date 2019 of December 2019 Denver has this call with Vizio and walks them through the problems explains it to them and then they ghost us. So Denver, a very very um uh persistent person. Denver

is um in some ways more persistent than me. And he kept sending emails. Every month he would send an email. I'm following up from our call. We were expecting a new source code candidate and they never replied. Now note that date. Now we're in the middle of the pandemic, the worst of the pandemic. Um, and then Denver goes on family leave uh for most of 2020 for

for most of the summer of 2021. When he went uh on family leave, he said, "Well, can you check in on some of the enforcement matters I had open? There's a, you know, there's a database here. Go poke around and if you have time, check in on them." So, the other thing was going on, of course, this was this was actually right during the second variant of

of like when we all got vaccines and then we thought we were going to be fine and then there was a new variant the vaccine didn't work for and then we're all like, you know, back in our houses again. And I was like on all these video calls and and video calls are horrible. Like I I miss old school like phone conference calls. And the reason why

is I've worked for home from home for about 15 years now. And the only reason I have folded laundry, clean kitchen, etc., etc. is because of conference calls, right? I would just walk around doing stuff while I was on mute and get stuff done in the house. On video, it's really hard to do that because they're expecting you to be staring at the stupid camera the whole

time. So, what I decided to do was what I decided to do. What? Yeah, I hate I hate using these kind of mics. I'm not good at it. Yeah, I hear I hear you. I'll do the best I can. It's just not the way I'm used to speaking. I'm sorry. Yeah, I look I'm doing the best I can here. We had we have like a series of

AV problems uh without support. So, I'm doing the best I can with the AV I've got. Um I I I I wanted a big screen, right? I wanted to I wanted to be able to like put a big screen in the kitchen, put one in the laundry area, and just do video call, like be able to walk around and do video calls. I'm like, what's the best

thing to use this for? A a Linux box with a giant display attached to it. Well, what is any TV today but a Linux box with a giant display attached to it? So, I was like, "You know what? I'm gonna go buy Denver wanted me to check in anyway. I'm gonna go buy a bunch of Vizio TVs um and see if maybe they just goes to Denver

because they fixed all the problems." So, so I tried them out and there was no offer for source code. Um, now it turns out if you read through all the litigation papers, you'll see they disputed that the the version of the TVs I bought, you could not see the offer for source until you connected the device to the internet, which is in our opinion a failure to

offer source code because you cannot assume the consumer who gets the device even can afford to have an internet connection. there are poor people in this world um who want to watch TV over the regular airways which still broadcast HGTV. Uh so uh so that was in our view a no source or offer violation. Uh we eventually conceded in the progress of the case that they had

addressed that issue. You now on current version at least well at least the error I last checked which is six months ago now you could see an offer for source even if the thing wasn't connected to the internet. So they have addressed that issue. So, they're making an offer, but the offer is no good. Um, and so we end up filing a lawsuit over the whole thing.

And we approach this lawsuit very differently than ones you might have heard about in the past. So, we are a copyright holder in a lot of different open source and free software programs. Busybox, Samba, Linux, a number of other uh programs. I don't know the whole list of all the copyrights we hold, but we did not file a copyright case. We did not You say it in

legal speak, you say we didn't plead the copyright holdings. What we did was we said, "Well, we're we bought these TVs. I went to Best Buy and Target and bought them for work because I was going to build these you video conference things in my house to do video conferences for work. So SFC is a consumer who purchased the device." So our argument in the case is

we are a third-party beneficiary of the GPL agreement and the LGPL agreement. What that means is the original developers copyrighted the software which allowed them to release in the GPL. Copyright's a really interesting thing, right? It's a it's a certain statutory right, but there's also this thing in the legal stuff. I'm not a lawyer. This is not legal advice, but this thing called the extra element test,

which says if you license your copyrights to somebody and then you make requirements, you still have a contractual claim if they don't meet those requirements. So, if I write a song for you, I can't write a song. So, if I write a poem for you, I can't write a poem. If I write um an essay for you on copy left, which I can do, and I decide

I've hate copy left now and I want to make it proprietary, and so I tell you, well, you if you want to redistribute this because copyright gives me the control to decide if you can redistribute it. Uh, you have to pay me a million dollars and I give it to you and you redistribute it. I now have two different claims against you. I have a claim of

copyright infringement because you infringe my copyrights because you distributed without permission. But I also have a contractual claim against you because you didn't pay me the million dollars. So, I can sue you both for copyright infringement and for contract uh breach or either. So in this case we only sued for the contractual breach but the contractual breach between the original developers who released the software under GPL

and LGPL and Vizio there they have a contract which is the GPL or LGPL depending on which program it is. So we are a third party beneficiary because that license says if you redistribute it you must make an offer for source valid to all third parties. that phrase literally appears in the GPL to give them source code. And the other thing we did is people ask a

lot like why didn't you just sue them for a lot of money? There's a really important reason we didn't do that. That's because we asked as our remedy not not money but specific performance. And what that means is we said specific performance of the complete corresponding source code that matches what the GPL and LGPL says that should be. So that was what our claim was. Now the

first thing Vizio did was claim that it was not a contract case and they got it what's called removed to federal court. So, what a defendant is allowed to do is say, "This claim has no right to be in state court. You should never have brought it in state court." And so, we're going to go to the federal court, and the federal court will actually grab the

case, pull it out of state court, and take it up to the federal level. Um, we obviously opposed that. There was a hearing uh emotions and all the usual legally things and the federal court for the Southern District of California decided that Vizio was wrong that in fact we had a contract claim the extra element test applied. It could be a pure contract claim and they remanded

it back to state court. So it went right back to the same court that it was in before. Now around this time we started to attempt to settle and we believe deeply uh that you should treat people as if they meant to act well and we felt we should give Vizio every opportunity to talk with us and try to get the source code correct. That's why we've

had so many rounds of candidates. We've had we've had three of four of them I think in the uh time of the case. That's after the other six we had way back in 2018 and 2019. Um and the other thing was going on is discovery. So, so discovery is this process in the legal system where each side gets to ask each other all sorts of questions. They

get to ask the people questions. So, they have to sit for things called depositions and they get to ask you interrogatories which are written questions and they just get to ask you for all the email you've written ever. Um I think I produced um I think 8,000 emails uh in the case. They they gave us an incredibly broad request and I was like I can search by

it and and and like don't don't say well where's that because I can't wave turn client privilege but there there was a thought of like is this going to be burdensome to produce and I was like no I got search tools I just search all my email forever so you know producing thousands and thousands of emails why they wanted all those emails I do not know they

asked me tons of questions about my personal relationship with Richard Stallman um and that it is no longer a good relationship u why they wanted to know that I don't know But uh apparently that's relevant to this case somehow. I don't know. They think it is. Um so that kind of stuff happens. But discovery process generally it's it's it's difficult but it's it's I love this quote

from uh Star Trek uh episode I forget what season it's called Measure of a Man. It's where Star Trek Next Generation where um they're trying to decide if data has rights. Um interesting topic for the day. Doesn't you does does an AI does an AI uh android have have human rights? And Pequard, lawyer Pequard is in his lawyer mode. He says, he says the a courtroom is

where is a crucible where we burn away irrelevancies to get to the truth. And I've always thought discovery process is that there's a lot of irrelevant stuff that comes up in discovery. And you burn away all the irrelevancies to get to the truth. So it's legal system's flawed, but it it has some there's motivations for why it's designed the way it is. Um, we amended the complaint

to concede. I mentioned before we conceded that there was, you know, there's now an offer. We're not trying to we're not trying to claim anymore that you don't make an offer because now you do. Um, so we amended the complaint and then we had there's all sorts of motions back and forth. They're all on a website. If you go to sfc sfconservy.orgvisio and there's a short URL

for sfc.govvisio, that will also get you there. You can read read everything or not everything. We put up basically key documents. Um, the court does a lot of procedural documents as well and we haven't put those up just because it's like you'll be clicking through and you'll be like the court affirms that everybody showed up and gave us their papers, you know, like that's not worth reading.

Um, so as I mentioned before, we showed up in January. Those of you here last year, I invited you all to come in September. We invited people online uh to come in January. I invite you again to come in August. Um, we do think this is pretty certain. watch our website obviously and don't book travel plans or if you book them do I mean you don't have

to book travel plans I guess you have to look at the traffic or something but uh if you're local but uh yeah don't uh make sure you check like the week weekend before the trial starts before you drive down to Orange County um so I want to talk a little bit about uh I think I'm going to like probably go right to time I have till 12:15

right yeah I think with all the AB stuff I think I think that's I think we lost time for questions because of all the AV problems here. So, I will be in the booth if you have questions, but I'm just going to get through the rest here. the there's a lot of GPL and LGPL source code uh that's still completely missing. There are binaries on the TV

for which we have absolutely no source code that are licensed on the GPL and LGPL. Some of the source code they produced has source code, but it is under an incompatible license with the GPL and or LGPL. And of course, we are still missing the scripts used to control compilation and installation. If you've seen my talks before, I've talked a lot about those scripts. And more importantly

related to those scripts, uh the LGPL programs do not operate uh with the modified version as the LGPL requires. Um you can actually get this source code candidate. Uh you would think because that's uh Microsoft's GitHub site, which usually what you find there is source code. You would think that would be a link to source code. It's not a link to a source code. they put a

document on GitHub that tells you how to request source code from them from another in a different method. I don't know why that's what they do. Um, so the list I'm going to give here, I'm just going to give you a rough list of some of the things that are still missing in the source code candidate. Um, I won't cover all of them and I'm not conceding

that this is the only ones. When when I'm played back at deposition this talk, which I may well be, or I'm sorry, at trial, I will remind everyone that I said this is not an exhaustive list. This is just a a selection of a few things to give you a flavor. So we're missing source code still for this for you know their this and their modifications to

E2FS progs to C library GUI library 2.2.21. They also have an copy of the GNU library 2.27 because they actually have a their firmware is a VM inside a VM. So it's very complex. So there's like two Yeah, it's very weird. Um and there's no source code for these modules at all. And the GPL specifically says you must provide the source code for the program in any

modules. And here is a list of modules for Linux which is GPL that were not provided. And not only that, many of these if you do mod info on them, it says license GPL. So Vizio has told us definitively that the some at least some of these are definitely GPL modules, but has provided no source code whatsoever for them. Now, here's the thing that Matt Matt like

like made a quizzical face about. Um, there are some of the source files. These are from Uboot, which is a GPL program where they provided a source file, but at the top of it, it starts with the information contained in here is confidential. This is a proprietary software. You may not copy, modify, or redistribute this software in any way. Um, so yeah, so they I I mean

this could be fixed with a with a wave of a pen, right? All they have to do is give us the source file again and say, "Oh, sorry. We meant to say it was GPL, but they have not done so. And then of course there's a scripts issue. Um they all these licenses require the scripts used to control compilation and installation of the executables and we've seen

nothing like that. And it's most apparent with the LGPL stuff because they seem to have never read section six of the LGPL because it's very clear about what you have to provide. And everyone kind of intuitively knows this, but very I I suspect many of you have not read this, at least recently. Um, the LGPL is designed to allow you to make a library that will be

known to be built with proprietary software. So it has all this extra language in it to say look when you recompile the library say the GNU C library either version that they provided or actually they're more than just two one of the many versions of the GNU C library they provided you have to give the user a mechanism so they can make changes to the C library

and then relink it somehow and run it with the binary that's there. And there's options for static and dynamic linking in different parts of section six. um whether you can use one or the other depends on various details but they have not done anything of this nature uh with regard to the requirements under 6A. So what's happening next? We may get lucky. We may get five minutes.

I talked fast. We may get five minutes for questions. Um keep that URL like bookmark it sfc.gov. Uh we'll update it. will let you know like please before you drive to Orange County for any hearing or trial you think is going on check the website before you leave the house just in case. Uh we don't think it's going to get rescheduled again but uh but we cannot

guarantee that. Um and we have really we've been queuing this up. It's funny I was desperately queuing up releasing various documents um that we were going to release uh from the discovery uh right before the trial just so people can look it over and we want people to have full context when the trial starts. Um, and so we're gonna we're going to release a lot of the

illrelevancies, illrelevant material from these big long depositions, which probably none of which are going to be discussed at trial, but we're going to as we get real close to trial, we're going to release a lot of that material. Um, so that you can if you do come to the trial or uh or just following the trial, you can you reference when they when they say, "Well, in

your deposition on page whatever," you'll hopefully have a copy of that. Um, some of the parts of the deposition are confidential, so we won't be able to release those, but most of them are non-confidential and can be released. Um, and there'll be the day of motions. So that's the Thursday before August 10th, whatever that is. I can't count right now because I'm so flustered from all the

A problems. But, uh, yeah, the Thursday before that Monday, there'll be a motion, which you can also come to if you just want to go Orange County for the day to hear. It'll probably just be like two hours. What's that? It's the sixth. Is it Thursday? Okay. Thank you. Thank you for doing math for me because I can't do math anymore. Um, so please come. Especially I

say this at scale every Just keep watching our website. Make sure it's happening. But a lot of you are locals, so it's it's easy to come down if you can get the time off from work. I understand if you can't, but uh we really want to make a good showing. Um the judge is already, I think, aware. We're not sure. I mean, the judge is, you know,

keeps everything close. It doesn't, you know, doesn't show any bias or anything. She's very good judge. Um but she she's she's clearly aware that there's interest in this case. And we want to make sure that everyone knows Vizio, the judge, us. we could use the support just the moral support of people being there that we know are there to support us and the case. So, if you

can get time off work and come down even just for some days of the trial, it would mean a lot to us to have to have people there. Um, and the funniest the funny part is I'm wondering if like we get enough people. There are only 20 seats in her gallery. And so, I'm hoping we fill it and they have to do an over a video overflow

room because that'll be really really cool. I'm almost to questions. I know you're desperately want to ask a question. Give me like 10 more seconds. Um, I have to thank Denver specifically. Uh, CC BISA requires me to do this. Uh, uh, that is a link to his Fosdom talk, which is, uh, since you, since my website's not up, you can go to Fosdom, search FOSM 2026, Deborra

Gingrich. Uh, this is a derivative work of that talk that he gave a couple of months ago. Uh, he gave me his slides. Um, so I was able to uh to uh to make a derivative work of his talk to give this talk. And I appreciate the work he did because it meant that I could truly do just in time talk prep this morning. Although I actually

should have just come here at 6 am to start working on the video, the A stuff. Okay, with that I will take this guy's got all these questions. Go ahead and give them the >> Thank you for your patience. I appreciate it. I'm very very flustered. I haven't been this flustered at a talk in years. I've never had this many ED >> All good. Um, August 1st

appears to be a Saturday. So the 3rd would be the >> August 10th. >> Oh, there was on my slide. I apologize. Got it. Thank you. Okay. 10th. >> The 10th is the trial. >> The 10th. Cool. I'm not a super open- source person, so a lot of this these concepts are new to me, but from what I'm understanding, you dumbing it down to folks who have

not been to any of your previous Vizio talks or understand the software conservancy coalition, whatever it's called. I >> Freedom Conservancy. >> Visio's using open source software. Therefore, everything they're doing on the device is supposed to be modifiable. So we should be able to reflash our Vizio TVs and do whatever we want with it. Is that an overstatement? >> A little bit oversimplified, right? So So their

base operating system is a Linux based system. >> So uh the parts that are under the GPL and LGPL which are called copy left license. So they're the ones that require what you just described. It is also permissible to aggregate proprietary software next side to that. So what Vizio did is they they built this you know this complex system that they have a Linux based system it's

actually it's actually one Linux based firmware and thenu running like in a true on top of that it's it's all comp very complicated but all that's open source software much of it is under the GP and LGPL and they do have to give the source code for that and do because it's Uboot and Linux the and gypscy and so forth they do have to give the ability

to replace those parts they're also totally allowed to have these aggregated proprietary things. In fact, I believe I'm not sure, but based on my analysis of their stuff, I think their entire TV app is just a web app. So, they have they have Chrome, they have Google Chrome, which is proprietary on the on the thing, and then they have their app. Um, and so that app is

proprietary and it would be totally permissible if they said, okay, if you want to uninstall the reinstall the GPL, LGPL, and so forth parts, none of that stuff's going to work anymore. you won't have a TV app anymore. It'll basically if if we get what we're asking for in the lawsuit, what we expect is we're going to boot the Vizio TV and we're going to have to

connect a USB keyboard or a Bluetooth keyboard to sit at a little bash prompt and that's all we'll have most likely. Uh and then from there you can install other stuff on it. My plan was as I said to install BBB or install my own web browser, install Firefox and then use BBB. So that so that's the thing. I'll get this next question then we'll jump back

to you if nobody else has anything. Go ahead. >> Cool. I was wondering whether it's just my perception that a lot of focus seems to be revolving around the LGPL rather than many other licenses that probably uh like were present or suffer under these licenses that allegedly or supposedly are present in in this firmware. Is it is there a focus on LGPL or am I just mis

is there just a misconception my perception? >> No, I I think I think there's there's two reasons there's a lot more interest in the LGPL than there used to be in GPL. Um, first of all, with regards spec I'll answer specifically to Vizio and then the general question. So, specifically to Vizio, I I don't know the exact count, but it's almost 50/50 I think of in our

list of accused programs. So, in the in the lawsuit we file complaint, we accuse that these programs, I think roughly half are LGPL and half are GPL. Um, so both licenses at issue in the case. Um, there's different issues with the licenses because they're obviously different licenses with different text in them. Um, and then in the general sense, I think there's been a bigger interest in the

LGPL, uh, and and people are choosing the LGPL2.1 over GPL 2 or GPL 3 a lot more. Um, because there's a lot of interesting things about the LGPL that aren't there in the GPL. Back in the old days, um I thought when I had we f I knew when I started like in the early 2000s doing this kind of work, um I kind of figured there was

someday going to be a really big case that went on a long time. It was like the most important case uh for GPL uh you know the watershed moment and I I think this may be that case. Uh, but I assumed it was going to be a derivative works case over somebody who had taken a GPL program, modified it, and tried to make it proprietary and claimed

it wasn't a derivative work. We were obsessed with the derivative work stuff or to use the GPL terminology, a work, a work based on the program. Um, in the end, that is generally not how people violate the GPL. That's not the clause. We thought that was the clause that companies were going to try to get away with violating. In the end, it's pretty rare um that that

that comes up. It's coming up here a little bit um with some of the programs, but uh I I think and I think people are realizing, well, if if most of what we're trying to protect against when we copy left to work is not that the LGPL has this flexibility in it that it does allow combining. And when you get to languages like Rust where where you

get one binary out and you need source code to build it all, the LGPLV21 is makes I I recommend LGL2.1 for Rust. I I don't I don't think GPLv2 or V3 is a good match for Rust programs. I think LGL2.1 I guess I'll take him again because he has the mic. So, just just one follow-up question on the the items that that you listed that were missing

source code. One of them was Dolby Vision. KO and the other one another one was uh HDR plus.ko. I would have to assume that Dolby Vision is not an open- source thing. Maybe HDR plus is. So, how do those twoepcepts of one of the ways that is displayed on the screen being perhaps open source or not, how does that relate to what you're describing as far as

which components are supposed to be released? Does that make sense? >> Yeah, they they they based so they're KO files, which means they're modules of Linux. They are based on Linux. In fact, some of them say if you do mod info on them, they say they're under the GPL. So if they say they're under the GPL and they combine with Linux, that is the work based on

a program thing that I was just talking with him about. So that's the one place where the workbased with a program thing does come up in this particular case. Um uh but yes that if and if they if they felt that they could and there's also section six of GPL um six and seven that kind of work together that say if if you if if you're in

a situation where um you are uh you have a proprietary component that's licensed in a different way and conditions are imposed on you by whoever you know whatever third party it is that says you may not distribute the source code under terms of the GPL, you are actually forbidden from distributing at all. So if they can't give us the source code of those modules, they aren't allowed

to distribute Linux at all. So that's their choice. That's what the license says. So hopefully that answers the question. I'm going to get his >> One last question, and you may not have a perspective here. You might have been thinking about it all over the last couple months is AI is just producing vast volumes of code and a huge amount of open sources in its training set.

But there's no human first of all, I think in many cases, paying attention to what that code is. Secondarily, that volume is just going to overwhelm the human legal resources to go chase it down. Kind of what does this all mean? And on the one way, it could bend the whole world towards open source in some very good ways. And on the other way, it's just chaos.

And if you have thoughts, I'd love to hear it. >> Yeah. In fact, if if I had known when the CFP closed that the trial was not going to be in January, um I would have pitched a talk that answers that question rather than this one because this is sort of regurgitiated somewhat from last year, which I apologize about. I've already apologized to the organizer about that

too. Um that's why I also added that the content which happened to be relevant this week. I hope that wasn't annoying to anybody that was like not as advertised, but I think that was useful and actually new. Um but yeah, I uh we we are working at SFC on um we we want to issue some statements about how to deal with free software and uh particularly LLM

AI coding agents which are really where it hits software. Um we did also publish this week. You should definitely take a look at the top blog post on our website um is an analysis of failor v pearl mutter which just on Monday so less than a week ago was uh denied certi denied cert by this uh by the supreme court of the US. So that means the

DC circuit decision that case stands and uh everybody was panicking because they thought it would you know that basically meant something with regard to the kind of things you were asking about and we have a good analysis there that sort of explains this is not this is not disaster for software and explains why. So you might want to look at that and I'm happy to talk in

the booth later. The other thing I want to ask is is it lunch now or is there another talk after me? >> There's another talk. >> Okay. All right. Well then we're done I think. >> Oh we have time. Okay. Go ahead. So, um, when you said you were going to talk about something recent that, uh, had just come up in California, I actually was not expecting

the age verification thing because like you, I had concluded that it was a non-issue. I was expecting you to talk about the 3D printer law that they're trying to put through. >> Yeah. In general, um, I'll just be really frank. We are an organization of of of six people total. Um, there are so many bills that get introduced in state and commonwealth legislatores around the US. We

do not have, we're not big and giant like the EFF, like you heard about this morning. I mean, the EFF has the resources to do that. Um, you should ask Cindy to give us a grant to do to do our stuff. I'm sure they could afford it. Um, uh, but but yeah, but because we don't have those resources, we generally don't p track pending bills. We just

can't. Uh, I mean, I could I I'm the only I'm the only person with an actual like my tal policy fellow. I'm the only person who's I mean I also have other duties that I'm trying to transition out of so I could be a full-time policy fellow but even once time full-time I alone couldn't track everything that's going to hit free software. So I am I vaguely

heard there was this law going through the legislature there's also a law another age verification law in Colorado uh there's something happened in Brazil with the age verification and open source. So, this is going on everywhere. Uh, and and we generally just have to wait until it looks urgent uh to really get in because I don't know if you f I have tracked state legislatores in the

past when things were going on uh like during back in the DMCA days and stuff uh because they were trying to push stuff the state legislatores that were is a mess, but trying to track state laws uh around the country is really really hard. And a lot of times people just your state sen state state congress people just get up there and introduce stuff. lot of state

legisl introduce a law. It ends up in committee and it's sitting there like like the age of education bill in Colorado sitting in committee like there's no reason for me to read that till it comes out of committee. Um but let alone I think has to pass their first their first uh their house before it goes to their senate to even bother. So that's why I don't

have an answer for you. It's a longwinded way of saying please donate to the software freedom conservancy. We're supported primarily by individual donations of $128 a year. 80% of our funding comes from that. Um, our government funding has been cut. Uh, um, we basically don't qualify for any federal grants. Now, if there are any federal grants, uh, I think you probably know why. Um, so we need

more than ever individuals to donate and give and support us. And so, if you're able to afford it, please do. Uh, if you're not, please promote um, you know, social media and so forth, uh, uh, if you like. I'm I'm on I forgot to put my social media up there. I'm atcopyleft.org. Um, and I'm easily trolled. So if you have questions for me, you you just be

buncopy.org. I'll probably answer unless you're mean. >> Well, with all the difficulty of the beginning, I think we had time for excellent questions. >> And and I want to thank all of you for your patience. I got so flustered. I' I've never had this much. I mean, I've been giving talks for 20 years. >> I've never had this many. I've had This is like the number one

now. There's number two that was almost this bad once. Um uh and I actually reschedled my talk that time. Uh, but yeah, I I got so flustered. I apologize was I was rude to you and Kurt and I apologize for that. It's it's very hard to be up here and especially when like my arm is in pain from holding this mic and I knew that would happen.

That's why I didn't want to use it. But thank you all for your patience. I'm glad we got through the whole talk and thank you for coming up and figuring out how to fix this thing when the AV people couldn't. >> I'll be in the booth. I'm head there And for the head thing, they said they had tape. >> I don't think I would have used tape.

I think I I think I just Sorry. >> No, it's not. You didn't do anything wrong. It's not your fault at all. I mean, I think I think it's I think >> should Move on to the mic. >> Sure. Test it, you know, put it on and test. Make sure sound is testing. >> Okay. Is this okay? All right. Cool. We'll see if it stays on. Check

check I think so too. >> Everyone hear me? Okay. So, all right. Cool. Awesome. Okay. I guess we can get started. So, hello everyone. Um, my name is Ray Peek for spending a beautiful Saturday uh with me inside. Uh, I mean I I was making a comment. I'm from the Bay Area and I mean even in the Bay Area like like Stephanie, you live pretty close to

where where I am. The weather's like pretty nice in March, but in Southern California it's even better. It's like morning it's warmer. Uh sometimes like in the morning you get up in the Bay Area it's like 48 degrees. And I mean, I know if you're from the Midwest, you're going to laugh chilly to us, but uh thanks for being here. Uh you probably noticed that my co-presenter

is not here. Uh unfortunately, uh this is a reflection of the like reality that we live today. Like he's based in uh Sweden. He didn't necessarily feel safe coming to the US. And I couldn't blame him. So uh so I'll be doing solo. uh so hopefully I'll be able to uh to do the uh the talk justice. I'll point out some of the observations that he shared

with me as we're building the stack and we actually presented this at Fosam about a month ago. Uh but it was a much shorter session about 20 minutes but uh the other thing like feel free to like if you have any questions during the presentations like you don't have to wait until the end just raise your hand. Uh I think all here he'll be he'll be helping

around with the microphone but you can shout it too and I'll repeat the questions so people know what the questions are uh on the recording. Uh a little bit about us. Uh like I said I'm in the Bay Area. I mean Fati and I have been working in open source for 10 plus years. Uh I mean I just means that we're old. Uh but we met each

other like when I joined the Linux Foundation in 2014. uh and Fati was one of the community members. He was working at Ericson at the time. So the the project has been renamed since then. Uh he was OPF was the name. I think the new name's Anuket under Linux Foundation networking and that's how we met each other. He was a community member. I was a community uh

I was a community manager. Um and uh you know at 2018 I left uh Linux Foundation to join GitLab. I was there for two years and I went to a company called Cubab and an semantic layer space and uh almost three years ago I joined my current employer Pinkap. Uh that's a company behind open source database called TIDB. Um with and then we have a booth here.

If you want to learn more about our product, come come swing by. And Fati, I mean, I think he did really well in the community under LF networking. He was voted as like a technical steering committee members and he was ultimately offered to come join the Linux foundation as executive director for uh CD Foundation. I mean, he has Fati has tremendous amount of experience in CI/CD. So,

he was definitely a good candidate. and then few years after that he he joined. He's currently a susay. Uh so both of us have experience working at foundations and commercial open source software companies. So that's uh that's why we wanted to uh put together this talk. Uh and we share a common hobby. Uh we obsess over it. And then Fati gets crazy like he goes berserk every

time I call soccer. But I'm I'm in the US so I don't have to feel very defensive about it. But first thing this morning I got up I was checking the FA Cup store. So that's how crazy I am about the sport. Um so I'm going to put a few logos here and I mean I'm I'm not like you know picking on these any of these companies

or communities but you probably look familiar. Uh a lot of these companies uh when they first started their company or project uh they started out with you know OSI approved open source licenses like Apache 2 uh typically uh and for various reason they decided to change their license to some variant of business source licenses and obviously there was a lot of uproar not just from the community

but open source purists in general right I mean there people were concerned that you know this is you know you're pulling the rug underneath uh community members uh and some companies like Cockroach and Source Grab like they continue to get bad press after the initial license change. I think it was like 2017 18 time frame. I apologize I forget the exact dates. Uh because they made changes

to like their core offering or you know further restricted visibility into the source code. So you know they'll get slammed again in the press and you know another discussion starts uh on on license changes. Uh so beyond you know license changes what's interesting is that uh community fork developed for for some of these projects like I mean some of these communities are here like valky and open

search uh so you know community members decided you know we're going to fork the project and we're going to put it under the opaces of I mean I think all three of these are under the Linux foundation they want it under a nonprofit like a neutral foundation um so which is which is great. Uh so you know community members like took action and they wanted a uh

you know they wanted their software project to continue under an open source license and what started happening when these foundations started getting formed was that a narrative started developing in like like online and the following exchange I apologize the text might be a little small but I'll I'll read through this so so you'll get the gist of the conversation. So conversations like this happen pretty frequently online

like whether it's Reddit or Hacker News. Uh somebody somebody in our community just made a quick post on Hacker News about hey this is TIDB a distributed uh SQL database that's open source that's written in Go and just linked to our repo. Uh and my colleague just chimed in with what few notes like here here are the main components of TIDB. you know, one of this is

is called TAKB, which is our storage layer that's been donated to the Linux Foundation, etc., etc. And then pretty predictably immediately afterwards, somebody chimed in and says, "Hey, this is a VCbacked company and it has a CLA, so this is evil like you know, where's a community for, right? When's it going to happen?" Uh, so I mean, that's fine. It's it's almost like predictable these like how

these discussions go online. And then somebody else, this is when I decided to post my reply. Somebody jumped in and said, "Hey, you know, I think the best model is like the Linux Foundation. This is be hosted under uh neutral nonprofit foundation and with multiple companies contributing." And this is when I jumped in and said, "Look, I work at both foundations at at and I I work

for Pinkap, but foundation isn't a panacea. I because I've seen you know from from the other side there are a lot of different there are a lot of different challenges. I'm not saying one's better than the other but you know grasses aren't always greener on the other side. Uh and this is when I got the idea of putting this uh presentation and this is when I when

I ping Fati and the that's how this this presentation idea came Um, so I'm going to just do a quick recap of how open- source foundations work. And I apologize some of this is is already familiar to you and you can uh you can tell me to stop if if I'm just re reiterating things that you already know. Um, so under foundations uh or all open source

projects needs to be funded somehow, right? whether it's you know company back open source project or even under foundations what happens is that most of the funding comes from membership like annual membership right and then you know in particular from like the top tier I'll call them platinum like the there are platinum level companies made significant investment into the project like tens of thousands of dollars is

how much uh the deal would cost each year and then not only is the amount significant for for platinum members when their projects typically begin, you're asked to make multi-year commitments. So there's a stable funding when the projects start uh because you don't want like a projects to be funded one year and then next year there's there's nothing available in case like platinum members drop out, right?

So they put those uh requirements in for platinum members. I think for CNCF it's actually like three years. So once you join as a platinum member, you're making a multi-year commitment uh so that you have a stable like a seed funding especially in the beginning so the project doesn't run out of money. Uh and so what do you do with all this money? Um so uh you

typically don't like I I mean I'll talk about this in a minute. you typically don't pay developers like like their salary from this funding, but you're you're doing paying for like the infrastructure. Uh you're paying for like foundation staff. I mean foundation you probably only have like five people like executive director, operations person, you know, one or two people in marketing for example. And then the other

expenses are like related to legal like you need somebody to look after copyrights and other IP related issues. And even if the lawyer is not on board full-time, like even if it's fractional, that's where the expense goes. Um, and and the other like uh typically like big projects like CNCF, a lot of their funding goes to like events like CubeCon. This it's not a trivial expense. It's

if you've been there, it's like a like it attracts 10,000 plus people. So it's not a cheap event to put together and it's important to bring community members together. And also it's important like a marketing vehicle for your project because that's how you get the awareness out uh get adoption and also hopefully attract like contributors to the project as well. Uh so that's where the lot of

the funding goes and I referenced briefly about uh technical contributions. So these member companies like especially platinum member companies not only bring dollars but they also bring significant developer resources. Uh so like using Fati as an example when he was at Ericson there was an understanding that he's spending X portion you know X percentage of his time working on OPNF right I mean so I mean it

depends on the company like for example there's an understanding that you know no one works 40 hours a week anymore but like 10 hours a week like 25% of your time needs it can be dedicated to OPNF or other open source projects that you're involved in. So that's where a lot of the bulk of the developer resources are coming from uh for these um you know foundation-based

projects and uh the companies that typically pay more this is not always the case but they tend to contribute most developer resources as well. Uh like I mean again there are exceptions like I've seen platinum companies that are happy to pay dollars but they don't necessarily contribute technically. Always thought it was weird. Uh and then there are silver level companies that punch above the weight in terms

of contributions, but by and large it's the platinum or top tier companies that are contributing most of the developer resources. and uh yeah and then you know another thing that that I often hear like see often online is you know once you put the foundation uh once you put the f project in a foundation like you're going to attract thousands of developers from day one. It's usually

not that simple. uh you typically I mean some a lot of times you see developers that are not from member companies contributing but the volume of those types of contributions actually pretty small in most cases from what I've seen I mean if there are if there are except like if if uh if there are exceptions I mean please let me know I'm not familiar with the cases

where most are coming from like non-member companies uh so that's what we typically see um and also obviously the governance is very different under a foundation like you have like you know governing board members typically from uh member companies but also like a technical steering committee is is is from like top contributors right um so I mean at some point you know even after like you know

when you launch a project you might have technical steering committees um sort of filled with like a platinum member like representatives but there's an expectation that after a year or two you need to open elections so that you elect like a technical leaders in the community. Um so the implication of that is is sometimes it's difficult to come to a consensus because you have a lot of

companies from uh you have different companies who are actually competitors part of a foundation and you know understandably sometimes difficult to come to a consensus right and then although you know according to bylaw you can just make a decision for example on the release cycle or approving a release through a majority vote You don't want to do something like a you don't want to decide on something

on a 9 to3 vote or 9 to4 vote because there's a significant minority that's not you know uh in agreement with the rest of the community. So uh even like putting something through a vote I you know that's as a foundation member we were pretty careful when we did that uh so that we want to make sure that there's a broad consensus uh so there's uh there's

not a there's not a disscent after the decisions been made by like a technical steering committee. So um yeah so one of the things I I want to stress here like it's it's easy enough these days especially these days to start a foundation like there are a lot of templates out there for governing laws and bylaws uh to put a new foundation and a project together but

what Fati and I are more concerned about because what I've seen from some of the foundation based projects is that what I'm more concerned about is sustainability after the first couple of years after the excitement and the hype dies down. Uh you know, especially after that two-year period is over, after the platinum member commitment is like done, like what if some members drop out? Uh significant numbers

of them drop down. And we've actually seen cases where I think one of the projects I've seen at the LF, they started, it was a relatively small project. They started with like a five platinum members and after two years like two out of the five founding platinum members like dropped out which obviously introduces a lot of gap in funding uh and it also is not a great

look either right so and it puts uh the project on a unstable ground after after just a couple of years. So that's something uh that's why like not just the foundation team members but the rest of the board they're pretty aggressive about recruiting like additional members to the project right because you don't want to run out of runway uh because you're running out of funding and people

are dropping out uh and as I mentioned a lot of contributors come from member companies because if members drop out that what typically happens is that the contributors can can be lost with the funding Uh so which which creates a pretty precarious situation. Um and you probably have seen this like especially big companies. The reason why some companies drop out of a project after a couple of

years is is many things like you know we we still see this like the company's decision to invest in open source projects. It comes in waves like sometimes there's plenty of interest and plenty of money available to invest in foundations but sometimes companies decides you know we need to cut cost or whatever the reason is they stop investing in it uh and I've seen this happen even

with like a big companies and and it introduces a lot of shock and uncertainty in the project right so um and let's see what was there was one more thing uh yeah I mean there's like the tagline line. This is this came from Fati. Uh it's easy enough to start a foundation, but what you need is like a share commitment long term uh to make sure that

there's a uh so that you know foundation doesn't provide that guarantee that it's going to be around after after a couple of years after the hype cycle dies down. So you need shared commitment and and then you know I think you need uh to be pretty you know aggressive in terms of being a lookout for I mean as a foundation staff I want to make sure that

you know you don't want to be surprised at the end of two years that somebody's dropping off right like you you need to know that ahead of time and then you need to be proactive in terms of making sure that you know existing members are are satis are happy but also you you're recruiting like new members into your project. All right. So, couple of examples. I mean,

one of this one is pretty extreme. You may be familiar with this. Uh just to prove a point that you know, foundation doesn't guarantee stability and and sustain and uh and stability. Uh so, the first one is the matrix foundation. Uh this blog post came out about a year ago. I think it was a couple of weeks after last year's vos. There's a block like at at

roads and can you like >> All right. All right. Is this better? >> All right. Cool. So um basically blog post says we're you know we've been so dependent on few companies for membership like we're our funding is starting to run dry and they had to defeature some of the free bridges they had uh for like a rocket sh and and few other platforms uh which I'm

sure was very painful right and then it was just a call to action was you know we need to recruit more member companies to to raise more funds um the good news is that I saw another blog post. This is like shortly before Christmas. I think on a financial front there things are moving in a better direction. I think it's not not all the funding gaps have

been addressed yet, but things are moving in the right direction. So hopefully things will continue to improve. And this is a foundation that's been around like I think over 10 years. Um so this is I mean this isn't like you know project that started two years ago and that are running out of stream but this is a longunning uh foundation but even they're having like difficulties which

which which is pretty tough to say. Um, so you know, like if just because you have foundations, they can't continue to like print money, right? They should they just need to have like more member organizations join uh to sort of top up the the funding needs that they have. Uh, and the second one, this is a more extreme example. I'm not sure if everyone's familiar with this.

Uh so there was a com company I hope it correctly uh Senadia uh they donated a project called nat.io to the CNCF um and I believe it was Apache 2. Uh but for whatever reason they decided that they didn't like that open source license anymore and they wanted to change the license. So, I mean, I think that's bad enough. Like, you know, company donating a project and

deciding to fork their own project under a different license, that's bad enough. But they went even further. They said, uh, cuz when you donate a project to a foundation, you're supposed to transfer your GitHub repo or or in most cases, you do. And also, the domain name gets transferred to the Linux Foundation in this case or or the CNCF. Um, so what Senadia said was they not

only wanted to change a license, but they wanted the domain name back transferred back to the company and same thing with the GitHub repo. The control they wanted to control the GitHub repo back to the company and obviously CNCF wasn't going to put up with that, right? So the blog post came out and and then you can you can you can Google it. Uh, and when I

I Fati actually texted me the link to the blog and when I read it, I read it like three times just to make sure I was reading it right because it was just it just seemed like completely bonkers, like how does this happen and how do how can this even end? Well, uh, and somewhat predictably, I think in about a week or two, they kind of came

to an agreement. So, the another blog post came out from CNCF saying is things are back to normal. We're we're back to where where things were. Um, you know, I agree this is a pretty extreme example, but you know, if somebody decides to do something against the wishes of the community, just because you're in at the under the foundation doesn't mean that you can stop it. So,

um, so it's kind of a cautionary tale and hopefully we won't see something like this again, but I thought it was really uh when I saw So did people when happened or like I must have been living under a rock because when I when I saw the blog post I said is this a joke and but it wasn't. so how do commercial companies work? Uh so I'm

going to I mean this is I think a lot of this is obvious so I'm going to talk about a few things here. Uh I mean people like to make fun of or be critical of VCs including myself. Uh but the VCs aren't the ones that are dictating how the companies are run. uh cuz you know obviously you get funding from the VCs but you have other

sources of funding called your revenue like whether you have a cloud business uh enterprise offering um you know consulting services I mean you have other you should have like if you're successful company you should have you know good revenue stream so you're not completely dependent on VC funding after a few years in business uh and another thing that I want to point out this is from a

CEO O at a company I was at before the LF and I mean we were an open source company but when we first got funding from a pretty big VC in Menlo Park uh during one of the open forums the CEO said just you know the board members are great they provide a lot of input they provide a lot of guidance make introductions besides giving us money

but the ultimately what the company decides to do is is up to the CEO he calls the shot the board can fire them but it's still like the the the CEO or or the company executive's decision. So I think we like to put like simplistically view and say, "Hey, company decided to change the license because the board told them to." Like I don't know if that's necessarily

the case. Like we're not in those board meetings. We're not even in the executive, you know, uh executive level meetings. So we don't know what what the re but I think it's like overly simplistic to say you know companies had to change a license because of the pressure from the board or the VCs like I'm not even sure if that's always the case in in in some

of the examples that I shared. Um of course like most of the contributions come from the employees um because you know company you know using their revenue or investment dollars to fund the engineering team right and then the expectation you're working on the project full-time like you know 9 to5 uh so of course like most of the contributions will come from the employees but you know like

like it is at like companies I worked at like you know gitlab or like today at t we welcome contrib contribution from from the community members, but the volumes are always going to be low because the contributors, you're not working full-time on on TIDB or whatever product the company is is working on. So, it's not like we're putting disadvantages on on the community members. It's just natural

that, you know, the employees of the company will have most of the contributions on GitHub or whatever other uh repos that you use. Um and what's also different obviously is that it's commercial companies is easier quicker they're quicker to make decisions because you're not relying on uh like a steering committee members uh comprised of people from different organizations. So it's easier and and faster to make decisions.

Um and community members may not always agree with it but you're not voiceless either. I mean if you're worth uh you know e even if you're not open source like I mean good companies should always listen to like a feedback from users or community members like although you know you're not going to listen you're not going to follow the advice of community members all the time but

uh you need to give them a fair hearing. So, you know, one of the things I usually tell people, uh, you know, starting when I was at GitLab is that as a community member, you have a voice. Like, you know, as long as you're constructive, don't be shy in in voicing it. And now, like more than ever, I think you have more like channels, sorry, uh, to

voice your opinion. So, don't be shy. Um, and like I said, as long as you're constructive, like there's no reason why a company won't hear what you have to say, and they'll be foolish not to, to be honest, so I wanted to also list like se like several companies or projects that are, you know, uh, started by commercial open source software companies and, you know, these are,

you know, very small set of communities that have not changed their licenses, right? or I don't know of any plans to and you know I'm going to jump to the last bullet here with the QR code. Uh so my my friend Don Foster she maintains a list of companies that have changed their licenses. This takes you to her GitHub repo. And if you look at the numbers

over the last 10 years, there's only been about 20 because when I f first saw that number because I thought the number was going to be higher than that because of all the negative press that gets generated once whenever somebody decides to change licenses. But 20 over 10 years isn't I mean you obviously you want to see zero. It's not like it's happening once a month, right?

So it's it's a pretty rare occurrence. Uh which is a good thing. And a lot of these companies are you know continue to commit to their open source principles which is good. So I think um you know a handful of like examples the past five years have sort of clouded our judgment sometimes but you know hopefully this will continue to be the case and most companies will

stay true to their open source print uh principles. And the other thing I out I mean since I worked at GitLab I mean I obviously have a direct personal um I think another important factor is like if the founders are still part of the company like their philosophy has a significant impact on the culture and the commitment to open source. Uh so you know one of the

co-founders of GitLab said I mean he he stepped down as a CEO but it was like a force of nature like if you do your work not in a transparent fashion or if you're not documenting everything so that community members and other employees know like what you're you know what your work processes are or how you organize hackathon you get a good talking to I mean not

just from from Sid but other employees too. So if you're not transparent, if you don't document your processes, that was a big no no. Um, and another good example, I don't know why I thought of this. Um, you know, a lot of internal GitLab meetings are like posted on YouTube. They have an unfiltered channel or they used to. So like you can hear like engineering meetings where

they talk about like prioritizing different features. Uh so so um so that's sort of the expectation that a lot of decision needs to be very transparent like even if there's somebody watches the discussion in the engineering team meeting after the fact and said for whatever reason I think he was having a one-on-one with one of the board members um and I think he just cut a snippet

like a seven minute clip of that discussion and it wasn't even about the product or technology it was about why you know Sid believes in like remote asynchronous work model and he was uh he couldn't understand why some companies have obsession having offices in in multiple places and he just you know took a clip of that and posted it on YouTube uh so people can uh see

what his thinking is in terms of like a remote asynchronous work so I mean he's a very unique individual so I don't you know expect all CEOs to behave like that but his you know philosophy and transparency uh is is critical to maintain the open source culture at GitLab and or and then I think he's continued to do that like he's like I think he's like a

uh he's one of the other companies he founded that you probably heard of it is Kilo code and they're behaving in a very similar fashion. Um so which is good to see. Um, so if you talk to like if you're evaluating open source like companies like commercial open source companies like I would encourage you to talk to the founders and see why they believe open source is

like important so and make sure that they're not just giving like platitude uh because some companies to be honest they do this just because they want to get more GitHub stars or increase like a download count for next round of investment. Uh but you know if somebody's genuine they're going to give a very different answer. Um so whether you're interviewing uh for a job or evaluating an

open source project talk to the founders or executives and you know get a good answer on why they believe open source is important uh for their future. All right. So the other thing I want to point out and I I referenced this early on uh you know many companies like commercial open source companies also donate some of their projects to foundations like you know Taikv I mentioned

already chaos mesh is another project uh that pink donated to the CNCF uh and three other examples like I think Docker was respon responsible for notary uh Kong with Kuma. So these companies, you know, not only have their own commercial open source project, but they donated some of their work to open source foundations. So they're very used to working in the open uh working under the foundation

model. Uh so they're not definitely not like a lot of these companies are definitely not averse to that. So uh that's another thing I wanted to really point out. And um um yeah, another thing I wanted to point out. Sorry, lost the point of So, um, so if you're using a commercial open source software or if you're a community member, how can you influence a company like

open source project? Uh, the first thing I want to point out is active engagement. And, you know, people typically think of like sending in pull requests uh on on the repository. I mean that's one uh but you know there are other ways you can contribute by sharing your expertise uh being an advocate for the company and the most important thing is is providing insight to how you're

using the product. Uh there was one engineering manager when I was at GitLab. I mean he was located in New Zealand but he was really prolific in in terms of reviewing community contributions because I mean it was part of my job to sort of triage all the uh uh merge requests that come in from the community. And I'll typically do it like twice a day. I'll do

it in the morning to catch the contributions that came overnight. you know, I'll do it towards the end of the day, Pacific time, to to review all the contribution that came on during the day, but even before I get to it, like he would like pick it up or he would assign it to somebody. And so I reached out to him to to first of all thank

him and and uh and uh you know why he enjoys like looking at community contributions. And what he told me was that it wasn't the code that he was interested in. He got real good insight into how people are using like particularly like new features because you know he told me like yeah I work with the product manager to develop a new feature but he doesn't have

a full sense of how people might actually use it or why they would want to use that feature. So through a merge request or even a like an issue somebody raised about hey can you like add this additional knobs to this so because we're using it in this environment. Uh so what Tuan told me was that what he appreciated more than code was the insight like cuz

he got a sense as to oh this is how they're actually using this feature or this is not helpful because it's missing XYZ knobs. Uh so that insight is very valuable. I mean code is great. I mean we're never going to say no to code but you know give providing us a visibility into what's helpful and and what's not quite working with the product that's even more

valuable. And there are a lot of different ways that you can voice your opinion. And like I said, you don't have to work like rely on the marketing resources of the company. Like if you want to, you know, create a blog post on LinkedIn, like you don't need anyone's approval, right? I mean, you don't have to go through, I mean, it's nice to get a blog post

posted on on the company's blog blog site, but you don't have to always rely on that. And you can, you know, have your own podcast or appear create a YouTube video. Uh so there are a lot of different ways to sort of express and your your insight uh into the product and the community. So I would definitely encourage a lot of people to do that. And as

I said earlier, I mean as long as you're constructive, don't feel, you know, defensive about reaching out to people at the company. Like if you're on the Slack channel, just ping the community member or whoever you think is appropriate. And uh and one of the things again I'm quoting said uh he would remind us that as an employee of GitLab, you're also a community member of that

project. So if another community member comes and wants to talk to you, you need to treat them with respect. So that's sort of the culture that he was setting. So that's what I try to remind myself too. I'm I'm also a community member and I need to treat everybody uh respectfully especially if you have a code of conduct or not only because you have a code of

conduct but that's the right thing to do and the other thing that I I mention to a lot of our open source users is that the value you get out of our software or the insight you have as an open source user is valuable to everybody including commercial customers. So your insight you share is, you know, it's not any less valuable because we're not getting any revenue

from you. So I wanted to share an example that we had uh just a couple of months ago. So Bling uh is a company in Brazil. It's an ERP company uh and they're expanding into other markets in uh Latin America as well. They're open source users and I noticed them on Slack because they've been they were posting a lot of questions. Uh but in addition to that,

they're also answering other people's questions that were coming up as people were like migrating to TIDB. So I I set up a call with them because I wanted to know who they are and and where they are in terms of their journey. And they were basically like migrating from MySQL to tidyb and and then their insight was like pretty interesting. Uh and when I met Casio, he's

he was the he's the infrastructure lead at bling. he came to our summit in Mountain View last year like uh my colleague Brian Foster who's a director of we encouraged them to write a blog post because I thought there how they did the evaluation of TIDB how they found TIDB their framework like was pretty helpful and that's going to be like applicable to whether you're an open

source user or a cloud user um so I mean this was like worth its weight in I mean Maybe weight and gold is not terminology to you analogy to use here, but it's like a really valuable asset in terms of our content. So, and they put a lot of work even on the blog. Uh, obviously like English is not their first language being from Brazil, but I

thought they did a fantastic job. And Brian and I, all I had to do was just some minor edits and uh uh and few suggestions here and there. But I mean this is you know a great way of sort of contributing to your project without writing any code. Um so I you know you know whether you're in a part of an open source foundation or or commercial

open source software community there are a lot of different ways to contribute and I encourage people to continue that. yeah, so some closing thoughts like I don't know what your impressions were when you saw the abstract or the title. Uh who should steer your project? And obviously it's one of those uh on the on the type of your project and and the nature of the project. Uh

it's kind of simplistic to say, hey, if it's not part of the foundation, it's just evil. Uh, and I don't think that's, you know, that's I mean, maybe I'm biased like I I realized as I was putting this presentation together, I spent more time last 10 years outside of the foundation rather than inside it. So maybe I'm biased, but I think it's just simplistic to put paint

a picture of this is, you know, what all commercial open source software looks like because all projects are different. And you know unfortunately Faty and I've seen examples uh of challenges of projects being under open source foundation and where things actually don't go very well in terms of sustainability. Uh because I worry about license. I mean it's you know our CTO Ed likes to say license is

not just a legal thing legal framework. It's a commitment you're making to your community members. So you you know changing the license is you know I mean maintaining your your commitment to the license is important but what I also care about is the sus sustainability of the project. So whether it's you know funded through foundation or company like I think the right model is going to be

very different uh depending on the project. Um and the final sort of thing that I want to stress uh you know we all have a role to play you know whether the project is based uh you know funded by a single company or a foundation we have a role to play in making sure that the project stays true to the open source uh ethos like I mean

and you you make sure that you know the whoever the leadership bar trace state to the principles by voicing any concerns or feedback. So we all have a role to play uh and rather than just throwing darts at a at the executives or or the board members like you know if we if we see any concerns we should voice our opinions so that they stay true to

the principles and also make sure help make sure that they have a sustainable future going forward. Uh so I think that's the last of what I had to say. Uh, thanks for listening, but we have plenty of time for questions. Thanks. Oh, here's um Hello. Hello. Can you hear me? >> Yeah. So, I'm Chris, uh, Norwegian government. Um, and we come in directly in the mic. Uh,

and we come at it from the fun side. So we have 40 projects super relevant talk for us in terms of where we're going now. We found it downstream. So these are projects that are reusing those components that you just mentioned open source frameworks >> that also is is maintained by the Linux Foundation. My question is when you talk about funding projects over time. So we have

found that none of our UN organizations, none of our projects actually none of them are members of the Linux Foundation even though they use many of the projects. >> Do you see that as an sort of as a as a base you could draw a foundation could draw funding for from or is it primarily the ones that are committing code and are close to the sort of

development process? Yeah. I mean, so like I'm not sure if I fully heard your question, but so where the most of the funding is coming from for for projects and foundations. Yeah. I mean, so I mean I think a lot of the examples that I mean because I was at the LF that I've seen is that uh when when projects form like because they have a like

a vested interest in making sure that they want to put a common infrastructure in the networking industry for example like or common implement a common standards. So, you know, they're motivated to do this because they it's it's by putting it under a foundation, you're externalizing your R&D. You're sharing your R&D like a resources among different companies because last thing you want to do is no doing the

same thing. Ericson the same thing in their silo and and Huawei doing the same thing in their silo. So they want to pull the resources together and and create a common standard but they'll find other different way to appreciate uh differentiate uh but yeah so that's what they're you know you know they're motivated to do this because you know they want to pull the resources to do

like uh joint R&D um so and then that's why they're contributing most of the developers right but hope that answers your your question sorry go ahead >> thank you for Thank you for the presentation. Sure. Um >> I um was interested by the dynamic of uh when you have an open source project managed by a company, most of the input into the project is going to be

um done through company staff, which you know um makes makes sense their full-time job versus um volunteer community. Um I was um I'm curious how you kind of prevent siloing when or just naturally like the things becoming insular to the company when most of the development is happening by a company staff. You mentioned um that the leadership and vision of the CEO is important for directing that,

but I was curious if there's any more um systemic uh things that you've seen companies take on to maintain that openness despite having so much of the development be um within the >> Right. Yeah. I mean that's a that's a very good question. I mean hopefully I think that's why I mean I'm trying to justify my own job. That's what a community manager is for. Like you're

like if you see that you know somebody made a contribution and no one's picking up, it's my job to go sort of find out who the right person is and and knock on the right doors. Um, and you know, my former boss actually at GitLab said this, like part of your job, I mean this, I'm not saying this to mean in a negative way, like part of

my job was to create some sort of attention like hey engineer like so and so I know you're busy implementing like X feature for for the next release, but there's this contribution that came from you need to carve out some time to take a look at this or at least respond, right? So that's um so I think you need a role whether it's a community management or

de devel developer advocate or even like engineering directors like you need to be sort of reviewing contributions that come and help sort of you know prioritize it or if it's not getting responded to at least like sort of escalate this um and then we're also looking at you know I was talking to our own engineering managers like everybody talks about AI I there are probably ways to

sort of automate this and like I I think like it's so long ago now like at GitLab a lot of the triaging was still manual but I've already seen after I left like things are just getting automated like it's getting able like a labels like automated like automatically applied and then assignment was done I was doing it manually but they were starting to do it like in

a uh more automatic fashion with the with this like a tool they had um so uh you know community members at least know that there's an activity right it's being assigned to somebody so they know who to like ping at you know at Pinkap or type or at at GitLab like who they need to like reach out to but is it perfect like not necessarily like I

don't think it's like a fully like I I don't think you can have a foolproof automation but you know so hopefully there's like somebody needs to sort of pick up the respon responsibility of making sure that uh things are not like getting getting buried especially in terms of community contributions and hope hopefully that answer answers your question but I think there's one more up front here so

yeah over here yeah nice shirt by the way give a shout out to Troy Dawson who has casual penguins is his story a lot of >> um but my question was so I'm kind of blending two parts of your topics you were talking about with influencing commercialled open source projects and kind of the um you know the ways like speaking up and getting involved with these projects

but then also you talked a little bit about with the nonprofit foundation model and you know one thing that's always been interesting to me is kind of also on the foundation side like how those people show up in the community because I've observed a lot of that work is kind of like back office operations kind of work which from a community perspective is often invisible you or

hard to So I'm just wondering like from a foundation model thinking about like the comp composition of this paid staff how do those people show up in the project community and what is the connection to the project community especially if all like the engineering is coming from outside of the foundation how do those people show up who are inside the >> yeah so I mean that's a

very good question I mean it actually you know I I hate it every time I I try to answer it like this it depends on the project like for us at OPNF We had a lot of IT staff uh because it was more of it was an integration project. So we had a lot of people doing like helping out with CI/CD, helping people out like FATI. Uh

and the way that we make those people visible is that you know you need to be part of like community meetings like even like a TSSE call so that they hear what's happening like a project wise and then they're also contributing to the discussions with the community members and so that needs to be made pretty transparent I think so you know so the community members know like

if I have a CI/CD issue like a heavy issue on Jenkins like who they need to call so that shouldn't be hidden uh But some projects I think actually CNCF is one of them like they don't have like a even a dedicated like IT budget when I last talked to them. Uh so it was kind of interesting like I don't know how that's done but with a

lot of the projects under LF networking like we had dedicated like uh IT staff uh and they would I mean one for one of the projects one of the LFIT staff member was even voted as member of the TSSC because he was making a lot of contributions. So uh yeah, you definitely don't want to hide any of the foundation people like even marketing people like they know

who like who's uh doing you know responsible for talking to press and analyst. But if you have a summit like all the staff members should show up I mean and then so that they can work with community members um and like having regular meetings even like a marketing committee meetings like uh I mean CNCF is a good example I I don't dial into all their monthly meetings

but you see all the foundation staff people there so you know who's work who's doing what from their marketing team. So hope that answers your question. Right. Cool. Right Stephanie. Yeah, I had a question about Oh, sorry. Is >> um so I'm working with a lot of uh projects that come out of academic settings. Um and a lot of them are like smaller projects that probably shouldn't

be their own like shouldn't be their own project and be integrated into like maybe a larger project. So I'm always trying to give them good advice on the best way to like approach it. >> Um a lot of them will go like, "Oh, I could just be part of the Linux Foundation." I'm like, "No, no, no. That's not how that works." You know, >> it's not cheap.

>> It's not cheap and it takes a while. >> Um, so I'm just trying to think of if have you seen like >> or are there factors that you would like suggest to people to when they're thinking about that next step? I know recognizing that it, you know, it kind of depends. Um, like if somebody's coming out building a community, when is it that they should be

looking towards these models or is it more that they should be looking towards existing projects to collaborate and and partner with and then be integrated into? >> Yeah. So, I'm not sure if I have a good answer for that, but but I hear you about like a lot of similar projects kind of perkle up and then is it's sometimes it's like an ego thing. They still want

to maintain their own turf and it's hard to sort of talk them into I mean this even happens under foundations by the way like if you look at all the portfolio projects under different foundations that seems like awfully familiar to the other website that I just read like two minutes ago. So for whatever reason like it's it's hard to like you know bring them together. Uh I

mean that sort of goes to my point about it's easy to start something start a project but you know in terms of like long-term maintaining it uh you know combining some of the projects if it makes sense and I mean I sort of went through that cycle in the LF like I started working on OPNFB but we came all under the elephant umbrella and that was actually

I mean part of it was driven uh by feedback from the member company because they were paying like different membership for different projects and and going to different board meetings and they were getting tired of that and I don't blame them right so but even after there was an agreement um you know at the board level like getting the different technical community members together was was actually

pretty pretty difficult I mean I I left less than a year after LFN was formed but it you know it's it's not it's not it's not easy right and then I mean one of the reasons why we did that was we we pulled the budget And then you know member companies were I think ultimately hopefully paying less yeah like I I don't know if I answered your

question in any way but yeah it's e easy to start stuff uh but if you want to maintain long-term sustainability of the project I mean that's the point that I I was pointing to at some point the right decision might be like a pull the resources together. That's what basically how LFN and other umbrella projects performed under the LF because having these like you know small like

a 15 small like projects under networking like really didn't make any sense right so it's it makes sense to pull the marketing resources pull the IT infrastructure even across different projects like makes more sense uh you know in theory that's how it should be but you know getting there is is not always like easy because people are not always happy about uh the decision to sort of

pull resources. Um so >> excellent questions anymore. Okay, thank you very much. >> Thank you. Thanks again for spending Saturday with us. Check, check, check. Everybody can hear >> No, it working. >> Can we hear Can we hear Kyle? >> Can you guys hear me? All right. one of us. I love that. Spread it all. Just spread it all out. Just checking if the auxiliary mic Where

am I? They'll show up. Don't worry about it. >> We're on time. Go ahead. It's now start. It's literally past six or whatever. >> All right. Check. Check one more time. Everybody can hear me. All right. I am super thrilled. I I'm actually ecstatic that you are all here. I am ecstatic that Basite is back again at scale. Just a >> I was told I whoop whooped

into the mic a little too much last year, so I'm going to try to try to not do that. Um, but I am really excited. I'm thrilled to be here to talk about Basite. Thrilled to be here with Kyle as well. And you know, I I guess I'll get us kicked off cuz it's it's time. So, this guy Kyle Gospo also, but a good friend of mine

and the founder, Basite. He's also a maintainer of Universal Blue and Frostyard. He's a Linux community engineer at Microsoft. No, I'm kidding. I love Microsoft. >> To say that. >> And he has been a player of Valve games since he was a young Wii lad. and uh and he's got some Basite swag. So, uh you know, mob him after the talk and it'll be funny. So, uh

and this here is my good friend Null. Uh I'm not sure why, but he's currently subjecting himself to Nick's OS, so it's a choice. Uh he's a core contributor at Basite and Universal Blue, and uh also our community manager, a technical account manager at Red Hat, and he very much enjoys spending time with his family and pets. I tried to do it. So, I'm going to start

off with um a raise of hands. Who here knows what Basite is? Has heard of Basite? Just kind of want to get it. All right, cool. I'll do the I'll do the shorter version, but still make sure I cover the basics at least >> most of the room. Thank you. >> Yes, thank you for being here. Yeah. So, Bazite really is a readytouse, convergent and bulletproof um

operating system. And it is meant to be ready to use to play your games and to basically work on any devices that you need it to work on. Whether it's desktops, laptops, handhelds, it does it all. It's fantastic. And the design in which that underpins Basite is why it is so great. It is bulletproof by design. It relies on our good friends upstream at Fedora and also

other tech, you know, and it's just awesome. And I and I love working on this project. And it's really just turn almost any device into a game console. That's really the goal. That's the design. That's the point of what we do. A big part in doing all of that is fostering a healthy community. Um we try our best to make sure that our contributors are happy, that

our users have a, you know, safe and inclusive experience at all times. And uh huge part of that as well is our very large team that you know helps us build all of this. Um fostering that community has allowed us to pull people from all over the place in Linux gaming including industry veterans. You can see the huge number of logos that we have here at the

bottom. So everybody from Ankor, Canonicle, Chainuard, Microsoft and Red Hat of course uh and also the cloudnative computing foundation. just really a huge swath of people. But having that large community means, you know, we have uh a lot of extra contributors to pull from. We get a lot of uh very unique perspectives on what Linux gaming should be. And a big part of our job is just

bringing them all uh together, allowing them to collaborate and allowing those uh disperate ideas to kind of meld into what Basite is today. Um the Linux Foundation provides uh very helpful project analytics that we have made use of. So this is our uh contributor health over the last uh quarter. Um we've gained 56 active uh contributors in that time and uh you know it shows us as

fairly healthy I would say. Uh this is mostly green. Um the organization dependency is just unfortunately me making a lot of noise but uh everything else looks you know good. I'm happy to see it. Um the last few months have just been this incredible uplift and interest uh amongst our contributors and we can't thank you all enough >> and that has led with all of this to

create basically an exponential growth for the project which is just unbelievable. I I cannot believe um the amount of people and good things that I hear about Basite, you know, on a regular basis from, you know, users, contributors, everything else. I mean, I've got people in my life, you know, who are just, you know, they've used Windows, whatever the case may be, or and they're like, I

didn't even realize that you work on Basite and I'm using it today. And I'm like, I feel like I told you about this, but maybe I didn't, you know, and it's just been such an exciting time to be part of the project and just to see this growth and trend and everything that that we do. So, I can't thank all of you enough. Honestly, I would, you

know, clap for everyone because it without you, we wouldn't be And our community is incredibly passionate. And I think that is like one of the coolest things about working on this project is the fact that our community really is everything. And my favorite photo honestly, I mean I love all of these photos by the way um from our contributors but and users but my favorite photo is

the top left one here where one of our contributors or not contributors I apologize one of our users um basically was setting up a gaming station for his daughter and he said the only thing I will put on there is basite and like I'm just like wow I about want to go cry because I have my kids of my And I'm just like, this is unbelievable. I

never would have thought that this was possible, you know, >> You know, it's amazing. I mean, that's a little girl gaming on Linux from the get-go. I don't think we could have said that 10 years ago. >> Um, so for our next section, we're going to cover some myths. Uh, we get a lot of, you know, different opinions about the model that's in use here and, you

know, see a lot of, uh, disparaging information about it. So, we want to try and break some of that up a little bit. Um, the three myths that we're going to cover are cloud native means your PC belongs to the cloud. Uh, some people, you know, see the word cloud and assume that means it's on a server somewhere. Uh, the second one is that you're limited in

package formats. Uh, this is more or less just covering um Linux packages. You know, things that you might use day-to-day that aren't in FlatHub. And the last one is that imagebased Linux is immutable. That word gets thrown around a lot in this space and I don't think it applies. So we'll cover why. So myth one um cloud native means your The thing about cloud native is it's

an approach to building. It's not a descriptor of where the software is actually ran. It just means that you are taking technologies that have been perfected in the Linux server space over the last couple decades and using them to build something in Linux. Um they're really common in servers. Of course, they're common on uh phones. Uh a lot of the same concepts are used even on Apple

computers. So it's widely more popular than traditional Linux desktops. Um our builds are done with just files which are built locally um or on GitHub, GitLab, Forge Joe. It really doesn't matter where you're building it. We choose to do it in GitHub, but it's just as capable on your local machine. There are uh 15.6 million cloudnative developers today. Uh that's a number provided by the cloudnative computing

foundation which surprise it's the Linux foundation. This is all just Linux. Why don't we just leverage those developers? I mean a lot of their previous contributor graph comes from people that were in the cloudnative space. It's a great place to pull talent and using that technology means everything that they already know applies here. Uh which brings me to the last point being industry standard with this cloudnative

tech means you have industry attention and resources. Um I can't thank GitHub enough for the amount of uh bandwidth that they've provided us in updates over the last year. It has to be in the uh high pabytes at this point. >> Cloudflare 2, >> Cloudflare 2, uh they're actually a sponsor. So thank you Cloudflare. But using those technologies has allowed us to do that meteoric growth that

no that null showed. And it's just incredible that we can even be here today. Uh myth two is that you're limited in package formats. So primary way that we handle packages on basite is uh flatub. Now flat hub is um a repository of flatpack applications that you know has also seen a very meteoric rise as of late. Um I think you know steamos and the steam deck

were part of that meteoric rise but um even just traditional Linux desktops are starting to adopt it. They've had 435 million downloads in 2025 alone, and they're having a over 20% year-over-year growth. I mean, it's incredible. Um, the other technology that we use is Brew for Linux. Uh, are any of you in the audience Mac users in your day-to-day or have you used homebrew on your Mac?

>> Yeah, quite a few of you. >> We promise we won't shame you. We we we do it, too. Don't worry. >> Absolutely. So um Brew supports Linux fully now and it is a fantastic way to get a lot of common CLI applications and even GUI applications. Um there are versions of VS Code and the um what's >> Jetrains >> JetBrain toolkit that are delivered over Brew

and they work flawlessly. It's awesome. A 100 million Apple users can't be wrong null. I mean at least a few of them you know >> and the last thing that we do to you know handle packages on these image- based uh Linux offerings is containers. Um the primary way we do that of course is Drobox but uh these containers allow for Fedora, Abuntu, Arch. Uh I've seen

people make Gen 2 images. Not sure why, but you can if you really want to. And all of those have just normal package managers that can install packages and export them. Um, if your package needs root access, you can make a rootful box and run them with root access. >> And here is just showing how simple it is to install something as a flatp pack. We are

so thankful um to the contributors of Bizaar, which is a project that basically its focus is to make it a seamless experience to be able to install and run your flatp packs. I literally just searched it like I do anything else. Android, iOS, whatever. Allows me to just search for the application, install it, and I'm off to the races. I've got heroic. And this is an example

of using Brew for Linux uh in two different ways. So, you see here, I tried to use the SL command, which was not found. Ran brew install SL. And there we go. I now have my fun Choo Choo train. Very important CLI applications on Brew. Now, in addition to that, we ship Bold Brew, which is a TUI version of Brew. So, here I'm just finding the SL

package I installed and removing it. We try to make this as simple as possible because the idea is not just developers are going to use Brew. Regular users will too. We want to make it seamless for And here's an example of using Drobox which um has historically been a CLI tool. But more and more love is coming from our friends you know cloudnative developers who are like

we need to make this easier for people. So I'm going to go ahead and install Arch because you know that is the best one. So, Arch, by the way, and I'm I'm going to create my little uh you know, my little base, but the the ability of being able to do common CLI options for my DRO box in order to be able to run, you know, the

software that I want to run and just make it seamlessly work, you know, anywhere that it can run is really, >> Got to go back here, null, >> just to point out that includes NVIDIA support right there with just a check box. This is one of my favorite Dro box features. It'll actually actually sim link the NVIDIA libraries on the host into the box and just use

them. >> So you don't have to think about it. It it simply works. >> It is so it is unbelievably slick. Like um you know whenever I am not able to find something in as as a flatp pack or I'm not able to find something as a brew package, I'm just like cool. All right. Dro box it up, pick something and just put it in there. It's

so so cool. So myth three is uh image-based Linux is immutable. I hear this word thrown around all the time and my opinion is that it doesn't really apply to what we're offering here. Uh immutable implies an unchanging route, something that is not modifiable at any point, even if you put the runtime descriptor on it. That's not necessarily true here. So with uh our model of imagebased

Linux, custom images can be created by forking a single GitHub repository and editing a container file or a build.sh file. You can throw arbitrary commands in there like dnf install or uh even said commands that just modify a single file on the disk that you want to change and that will spit out a perfectly functioning image that you can then rebase to. It's the basically pseudo version

of pseudo in this offering where you just have full control of the image and no one can tell you otherwise. Option two is that uh we have RPM OS tree allowing RPM packages to be layered on top of the system and applied either live with a command or on the very next reboot which is helpful because if you layer something that is broken for some reason upstream

like a common Linux issue is uh dependency issues the image will either not build and the command will simply fail because it could not resolve the dependencies or if something's broken past dependency resolution the image will succeed. You will reboot. You'll find that your new image failed to boot and you can simply select the previous image as if nothing went wrong in the first place. That's the

bulletproof at work. It's awesome. >> Yeah. Uh we also offer systemd system extensions which are relatively a new feature of the systemd suite of utilities. Um, these are disk images that contain a modified version of your root, just the changes that you want to make that are layered on top of your running system. And these do not require a reboot to work. So, uh, friend of mine

at Red Hat, his name's uh, Timothy, he built a repository that allows the creation of these system extensions from Fedora RPMs, which is very impressive. I mean, even versions of Steam have been packaged in this way. So you can just convert an RPM over into a system extension. Drag and drop it into a folder and you now have your system extension live and loaded. Uh final point

is just immutable implies an unchanging readon root file system. After all this, I don't think it describes what's on offer here. This is simply modern Linux. It's image-based Linux. It's different and it requires learning just a little bit of that something different which we all know Linux users don't like to do. It's so hard. And this is just a basic example of um doing exactly that. Editing

our image template and installing one of the most >> the most important package in Fedora, which is um doing uh doing some doing some cows. And on top of that, you can get something like this, which um I might be over at the Doghouse tonight. I definitely was on there Thursday night when I flew in. I love that place. If you have not tried it, it's absolutely

wonderful and amazing. But I suppose if you do want to learn about how to actually do this, you can look at the image template and you know, it is what it is. You know, like >> please scan the QR code for the image template. We have a dog house on the right too if you're interested in going there. Depending on how game night goes, you can probably

find me and whatever's left of null there tonight. Uh just covering briefly what I was mentioning about system extensions. Um we had all these fun videos earlier that showed you know all us installing something or layering something. Now this is a still image here. This is a systemd system extension in the var lib extensions folder. Installing them is this simple. It's drag and drop. I don't even

have to show you the action of doing it. So, it's a very powerful tool. Um, I think with the work Timothy is doing, it's going to probably be the future of layering packages on Fedora rather than RPMos tree, but we'll see how things develop. And this is an example of another tool that Timothy Ravier is working on, which love the guy to pieces, he's so great. And

this tool basically is a CISX manager tool which allows you to have basically a simpler way of managing your existing um systems extensions. And so basically you make the installation, you go ahead and um you can check the status to see what actual CIS X you have. Um it actually runs as a CIS X. Um, oh >> yeah, trying to pause that for you. I guess it's

a GIF. You can't. >> Yeah, no, that's okay. But essentially, it it is a easier way of being able to um manage these things from the command line. And ultimately, the goal with, you know, reading what he he posted on here is that this is a proof of concept to basically merge eventually upstream so that it's easier just in general to be able to manage this. So

when we're talking about the future, which is, you know, what was K Kyle was saying earlier, it's it's a really bright future. It's very cool. I've played with CISX. It just works like he said, drag and drop. And but this is a way of, you know, so he's showing an example of installing tree and just adding it from the extensions repo for Fedora core OS. He's doing

an update to basically make sure that it's actually there. And then you can see the status and see cool now I have the tree command which I didn't have before. And so he's going to go ahead and enable it. And then he's gonna see the folder which is what Kyle was talking about where you can just literally drag and drop the file but it looks like it

has a um sim link there. Going to go ahead and refresh. Make sure it's you know actually showing up and then um check. Cool. We've got it. And then we're going to go ahead and run the tree command. And there you go. But it can do more than just SL or tree. It can actually, you know, install even more useful applications. In this case, we want GDB

on there. So, we want to do just some debugging. So, cool. We're going to download the update. Go ahead. Cool. Now, now we'll just go through the same process again of enabling it. And then we have, you know, we can disable tree and get rid of it if we no longer need it. do a refresh and there you go. You've got GDB. >> it's pretty slick. I'm

>> It just works. >> I think a cool point here too is much like the uh app image manager of the past where the app image manager to manage your app images was itself an app image. It's a miracle I was able to say that without any uh stumbles. Um this is actually a system extension. The first part of the script is just building and installing the

system extension manager Now uh this is an example of using RPM OS tree to layer something today. I know we recommend to our users that they should not do that but in this case it's AMDL which is a command line utility for undervolting AMD CPUs. It's not something that we would ever pre-install for a user and it's also not something likely to have any kind of uh

dependency issue going forward. its actual needs are pretty light and it needs root access. So doing a dro box doesn't make much sense. Um you can move on to the next one. It's just simply installing and rebooting. I think now at this point we can talk a little bit about the Linux gaming ecosystem of today. Kind of a fun topic because there's a lot going on there

right now. Linux gaming has a number of issues that affect it today. uh primarily it's uh fragmentation uh common dependencies are you know not well documented there's a lot of uh tribal knowledges with little tweaks and udev rules and drivers and all sorts of other things that all these different uh offerings are shipping that you know don't get communicated or kind of get brought to other offerings

peacemeal um that institutional knowledge is not which you know it leads to it leads to complications it leads to strife. The last one is a very common ancient Linux problem which is distribution packaging. You might have a gaming image that's or gaming Linux distro that's on uh Arch which has Arch packages. You might have one on Debian. You might have one on Fedora. Uh there might one

be B one based on a DRO that doesn't yet exist today. Each one of them has its own packaging requirements and its own formats and its own necessary tweaks. And if one is packaging something brand new for their DRO, they're probably not packaging it for yours. >> But but mine is the right one. Oh, >> always every time. >> Yeah. Yeah. My my standard is going to

be the one that kills all the other >> Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. >> Another problem that's affecting us today is hardware support. Uh picture not related. A lot of hardware vendors ship closed source blobs, which more or less means there's nothing the community can do to fix problems that arise from those blobs. Um, avoiding that really requires voting with your wallet because the community can

do very little except build something completely different that reverse engineers or makes that hardware work without the blob where possible. Another problem is uh multiple vendors. Um I don't want to single ASUS out here because they've been awesome. They're phenomenal, but the original Ally shipped a fingerprint sensor. Super cool that the hardware had it, but they used two different vendors when they shipped that sensor. One vendor

had full Linux driver support out of the box, which of course was the one that I got when I tested it and was like, "Oh, everything works great." One vendor had no Linux driver support and still doesn't to this day. So, it's kind of a lottery system which piece of hardware that you get. You might have a working fingerprint sensor, you might not. And the very last

one is manufacturer in uh interest. At this point now, Linux users are about 5% of the desktop market roughly. They're a little less the gaming market. And while a considerable number, it is still a number that could be written off if it means a lot of extra work has to be done on the manufacturer side. So the best thing that you can do as an audience is

to buy hardware that already works out of the box today and support the people that make that effort because they need a reason to continue to make that effort. That's for sure. And as Linux grows, all of these problems will only get better. >> And encourage your friends and family too because that's like another big thing as well is encouraging others in your community to support the

things that support the things you care about. I mean, I know we all do that, but it's so important. Another hot button topic in Linux gaming is, of course, anti-che. Um, I think the screenshot is about as old as me. It definitely looks crusty enough, but antiche is a game industry problem. It is not a Linux problem. Uh, in my past life, I worked on e-commerce websites.

And if you talk to anyone that has to work in servers anywhere, they're probably going to tell you to never trust the client. It's it's rule number one of just anything secure. So what has happened with the game industry is they've decided, okay, I can trust the client if I have a kernel mode driver running on their machine. And the answer is you really can't. Uh I

can have external hardware outside the machine that watches my screen and moves the mouse and does the same thing. The client is never trustworthy. So, the proper fix would be for modifying anti-che to run server side. Um, games like Minecraft actually do that today and Valve's latest version of VAC does that. So, props to them. They're doing the right thing. I I adore that work. Uh, and

that just means making sure that the client is never sent packets for, you know, information on where players are. They can't see. They never get um they can never decide how fast their own character is moving. it's limited by the server checking the actual you know capabilities of the game logic and other simple server side checks like that. Uh more complicated ones are checking for um aim

hacks which there have been some uh people that have had good success with using heristics to detect that because a player that can just snap from one person's head to the other doesn't aim like a normal player and that is expressed in the packets that they are sending to the server. >> Except for me I'm really good. >> Good gamer chair. Um, same answers before. Of course,

vote with your wallet. You know, buy games that support Linux antiche today. Uh, there's Valvac I mentioned works. Epics EAC has support. Battle Eye has support. Punkbuster has support. It's up to the game developers themselves to turn on that support and allow it to work on Linux. It's not Linux's issue preventing Make them work for your money. This is an economic problem and not a technical Uh,

and speaking of voting with your wallet, I think it's just worth a quick shout out. Stop Killing Games initiative. Uh, they've ended here in the US, but the European one is still going on. Their goal is to prevent, uh, publishers from turning off games that you've purchased. It doesn't mean they have to support it indefinitely. It just means there has to be an out where the community

can take over that support. Run an old server or um, you know, release partial source code so that something could be maintained. however they choose to do it. Just keep your games that you bought working. >> And I want to make a call back to a slide that we had last year, which was um man, this art is cool. I I I was like, Kyle, we have

to use this slide again. Like I I can't do the presentation without doing it. And really Linux, you know, as you know, a community sits on the shoulders of giants, sits, you know, on on top of other projects and really works hard to basically make things better and better over time. And I think like the growth that we've had as a community, just all the interests around

it is it's it's it's amazing. like we should be so proud of ourselves as the community and I am just like so thankful to be working with so many cool people um in the industry that my goodness I never thought I would actually ever like ever be able to do we decided to do co-op mode because we can't do this alone we as you know the Linux

gaming community and you a bunch of projects have gotten together and have said enough is enough. We're going to share this stuff and because we've already been doing it anyway. That's the real secret is the fact that we have been sharing this stuff. We've been sharing patches and what we want is we want the industry to start paying attention and saying here's the place you go >>

and a lot of them already are. It's more about providing them a centralized place where hey you come talk to this this entity and you get everyone else >> Exactly. Yep. It is it is the place to go and the intention is to work together to get these things upstream so that everybody benefits not just one singular >> Yeah. Um we have some OGC members in the

audience. If you're willing we'd love it if you stood up and even came up here you know we can be part of the questioning aspect of this. Clap. >> I love these guys. >> Thank you you guys. >> I'm really glad you're here. >> The other thing is we need your help because it can't just be us alone. Like I said, we've got lots of work to

do in order to make gaming even better than it already is. And I am so thrilled to, you know, just put out this call to action and just say, "Hey, let's get this going." Like, no matter what distribution you're on, you can benefit from doing this. No matter what, you know, yeah, RPMs are the best, quote unquote, or Nyx is the best or whatever the case may

be. But we're trying to make this so that everybody can game on Linux because that's what I want. Everybody should be able to do that. the too long didn't watch of this talk for those of you who are what clicking on the video in our discord or something um cloudnative Linux or imagebased Linux is anything but immutable um no I will not elaborate being approachable to new

users doesn't mean you can't make overarching changes to your install of course we allow uh arbitrary changes to the image vote with your wallet um Linux has a considerable user base and it can drive real change as as long as everyone is making decisions that benefit Linux. Um, the only fragmentation you want in your games is uh the fragmentation grenade. So, let's try to keep the the

fragmentation in the Linux community out of the Linux gaming space. We want to support everyone equally. Uh, the last thing is working together means getting more for less. It benefits everybody. So, always always and I don't have to say this to everyone here at scale. I mean, we're all friends here. Every every dro on on Earth is here right now. So, thank you all for being here.

I'm broadcast. >> Sure. Thank you. >> So, now is the section for the Q&A. >> Ask anything you want. We're all here. >> This is an OGT Q&A now. It doesn't have to be Basite related. >> Hello. Okay. I' I've recently tried just basic, you know, regular Fedora and I'm coming to learn about SE Linux. Do I have to worry about SE Linux in in here >>

Fedora does a very good job with their default SE Linux config? Um, there is a single issue that we've had with it in Basite's entire existence which is related to old 32-bit Valve games and their use of a audio library that needs exec permissions. Most of the time you're not going to run into it. If you do, um, whatever tweak you would have had to have made

in Fedora is the same tweak here. It's one to one. And you can always turn it off too. I don't recommend that, but you can. Hi. >> Okay. Um, you said you're looking for help wanted. What type of projects? I've got some CIS students who might need some honors projects. >> Oh, that's really cool. Anyone want to cover anything they're interested in? >> Um, hi. I'm one

of the Fear Labs people. We're in OGC. We're always looking for RPM packages for the Terror repository that Basite and Ultramarine Linux share. So, if you want to Yeah. What do you say? too. >> No is might use it. You can use >> it. >> Yeah, Eggy. Let's go. Um, but if your students want to learn RPM packaging, uh, we have a great guide on how to

do it. So, developer.farolabs.com. We'll have links somewhere at some >> And ask me and I'll help you. >> Ask this guy. >> I I love RPM packaging a lot. Um, I maintain like 17th of the Terra packages that we have and I love helping new contributors. anything from a open gamepad UI perspective? >> No. >> Okay, that's fair. >> The answer is we're always looking. So, um,

if you look in the Basite GitHub or the Open Gaming Collective GitHub repository, any any repo in there that can take any change at all, we're accepting pull requests. We'll be happy to have them. And of course, if it's a project too that's just potentially something beneficial Linux gaming, reach out to us because we might just want to pull that into the OGC and provide extra support

for it as well. >> Yeah, I was wondering about um I guess kind of like the focus as you go forward like it seems to be like the focus right now is kind of like the enablement of gaming on Linux and gaming as like a consumer as like the end gamer, but first OS and like are there any ways that we can get involved to help um

be cool if I talk about the images that we're thinking? Yeah. All right. So, there's uh two ways that we're thinking about that. Um from Basite's perspective, we're working on something called uh Basite GDX. Uh already today, there's a version of Basite called Basite DX, which is the bluefin developer experience shipping on top of Basite. Um, it just means you have VS Code and a number of

other tools pre-installed and ready to use. The GDX version is going to expand that so that you have, you know, 3D artistry tools. You've got access to things like GDAU or Unity or the Unreal Engine and, you know, whatever else a game developer might need. So right now we're kind of accepting comments on that and looking for what would be best to offer or most helpful because

the thinking is this is a not necessarily reproducible but this is a widely deployable image where every piece of hardware you have has the same experience so it could be useful in a studio. Um, from the OGC perspective, you know, if you're not necessarily a game developer, but you're a Linux developer, we're looking at leveraging this uh, imagebased technology to ship reference images of different distros with

the OGC toolkits put on top of it. So a version of Debian with the OGC kernel and open gamepad UI version of Arch a version of Pedora not for end user use but for other Linux projects in the community to say you know oh this container file has the exact steps needed to turn Debian into a gaming dro I'll take this and bring it into my project

and that's something we're going to use to support as many other projects as we can. Uh could you briefly talk about uh what your approach is for um experimental stuff and and preview stuff? I I recently started using the heroic games launcher for my GOG games because I saw something saying that it supported achievements. Turns out it's not really in the main branch yet and stuff. So

>> I see. Um well from Basite's perspective, it just means that you might need to pull it from a different, you know, repo. Um, if you were trying to use a feature that wasn't in the flatub release, which is typically how heroics installed on basite and you know typically shipping stable, then you could always switch over to Dro Box and use a nightly version of heroic that

might have that feature turned on. >> You can also layer it if you want. >> You can also layer it with our PMOS. We have heroic and terra. You should use the flat pack, but that's an option. >> Take it. Take it. So, uh, I'm part of another group that we work with, uh, Heroic and Lutris and all the game specific launchers. And, uh, with Heroic in

particular, I would definitely strongly advise using the flat pack because they keep that mostly up to date and that's where a lot of their main attention comes from. Um, the other issue is when you get to packages that are provided by distribution or a third party repository as opposed to the flat pack, then you get into Well, maybe they may not maintain it directly or they may

or the distribution is behind in their versioning. So again, stressing a flat >> I think it's great though that you know all these different Linux projects or you know friends or at least talk to each other. Linux in general it's it's >> So I had a question about hardware support in imagebased OSS. Um, my motherboard has a weird fan sensor that needs an outofree kernel driver to

be able to actually work. And it needs to be force installed in a way that doesn't necessarily make sense to have everybody do. And in a non-image based thing, I wouldn't need to fork an image and then do all of this in a forked image and then maintain that fork of an image. I would just have something installed and be able to make sure that I update

that kernel driver every time I update my kernel. And so I'm wondering from a UX perspective for people with slightly tricky hardware, is there going to be an answer better than fork the image? Because I wouldn't want to tell a friend that isn't as into this stuff as me that is trying to get into Basite on their machine that oh, I'll just have to fork this image

for you and do this weird thing so that way you can actually control your fans. >> Yeah, the I mean the unfortunate answer is that's just a limitation of image based. there's nothing really coming to fix it. And right now building an outofree kernel module for imagebased means you have a separate repository that has the exact kernel in it, builds the AK mod for the kernel and

then gets packages and RPM to be installed into the final image and it just does that every image build. Um, part of what OGC is trying to do is uplift these projects that you know are out of tree but need to be in tree. try to get them to actually submit it or if they have submitted, you know, get them the support that they need to go

get it finalized and approved and in the Linux kernel. And part of it is going to be voting with your wallet. I if you already own the hardware, it's it's awful and I'm sorry, but if you can't avoid buying that hardware in the first place, that's the best way to make sure it works. >> Oh, yeah. Take >> Yeah. So, so the real answer to that question

is um figure out why that kernel module can't be upstreamed and get that fixed so it can be upstreamed, right? That's a intensive process and that's part of the reason that OGC exists is so that we can help guide that that pathway forward, right? And in the meantime, right? So, we have an upstream first mentality with our kernel. So, we want to push that upstream, but it

can take quite a while to get those approved. So we can also ship that kernel driver in our kernel ahead of time while it gets reviewed. You know, if they takes, you know, you added that semicolon when you didn't need to. That's like a two week turnaround, you >> Yeah. But yeah, that's the goal is to take those like niche things and get them upstream so we

don't have to worry about like using AK mods or TKMS >> Yeah, I think that's a great point that you brought up too. Um, part of, you know, help wanted is it goes both ways. If you're working on a Linux driver for some unique piece of gaming hardware and you need testers, come talk to the OGC. As long as you're working on uh upstreaming it and it

doesn't horribly break things, that's more than 70,000 plus weekly users that can test your driver and make sure it gets into Linux kernel. >> Very good. Any >> any other questions? >> Any more questions? Uh, so we'll thank the presenter. >> Wonderful. >> Thank you all for being here. >> And thank you guys for being here, too. Seriously. >> Come on. >> Good job, you guys. Thank

you. >> We're I think we're a little past even. We're done. >> Get the Get the Get it out. >> The next >> We'll see. >> It's all right. Hey, let it go. We'll We'll see you guys at game night. >> We'll let you free. up pretty soon. >> there we go. See, one more minute. Yeah, turn those off. The other thing you could be picking up

is this little microphone up here. This black one. It might be picking it up from that, too. I don't know. Okay, I can do this and then I'm going to do this. This is the easiest way for me to reset Uh oh. What happened here? Oh, there we go. Hello ladies and gentlemen. My name is John Mad Dog Hall and I am the board chair emmeritus of

the Linux Professional Institute. That and $350 will get you a cup of coffee at Starbucks almost any place. Unless of course they're closed like at 2:00. Ridiculous. I didn't want to I want to announce the old farts buff. We kind of got a late announcement of this, but it is going to be held in room 209 from 1900 to 2000 hours tonight. And uh if there's any

Windows users in the room, that's 7:00 p.m. to 8:00 p.m. So, for all of you who wonder who this weird old guy, this old fart who's standing up here on the stage, who is Mad Dog? Well, I have been in the computer industry for a long time, you know, and I'm not going to go through all the junk on here. The most important thing, of course, is

the Dilbert slide up at the Dilbert cartoon up in the corner where if you want to, here's a nickel. Go get yourself a real operating Now, I started writing programs in 1969. We didn't call them open source. We mainly distribute them in source code because there wasn't any sense in creating binary images and sending them around because there weren't that many computers around. Okay. Um and I

met Richard Stallman in 1986. I'd gone to work for Digital Equipment Corporation and we wanted to help out the GNU Foundation putting their code or the FSF actually putting the GNU code onto Altrix our system. and I met him by giving him his first Vax micro computer. In May of 1994, I met Lenus Torva, this person I'd never heard of who was working on a project I

didn't know anything about. I was asked to fund his trip to Dikas in New Orleans and uh it was there I saw Linux for the first working for the first time and said this is the best thing since sliced bread. I've been working on a project called Canina's Lucas down in Brazil ever since 2012. This was to help people in Brazil has a very high import duty

rate to be able to get a system for a reasonable price. But more than that, it was to help them learn how to design and manufacture and create supply chains and support free software and open hardware. And we'll get more into that later on. And I've been a Bernie boy since 2015. Unabashed, flaming liberal. I am left of Bernie. Actually, I've been to over 100 countries, more

most of them more than one time. I am married to a Brazilian and we just celebrated our ninth wedding anniversary last Saturday. Now, I'm going to remind you of what happened before capitalism. In cap before capitalism, there was surfom. Okay? There were kings and there were serfs and the kings told the serfs what to do, when to marry, when to have children and stuff like that. Later

on, the kings kind of shared that power with the church. Okay, but for you know the power still came from the king and they made and executed and I use that term strongly their laws and the surfs basically owned nothing. Now in the United States, we kind of overcame King George, but we started a different one. And we had masters and slaves, okay? And again, the masters

owned everything and the slaves owned nothing. And they told the slaves what to do, when to marry, what to eat, when to have kids. They sold off their kids. They were property. And the slaves owned nothing. Now we have closed source companies. They are the masters. They're the ones that tell you where you can put your software, how many people can use your software, when you can

upgrade your software, when you have to throw away your hardware because the software no longer works on it, Windows 11, and when to retire their software. Okay? And I will remind you that there's about 10 million people that are still using Windows XP for various reasons, okay? And they can't get patches and they can't get put onto new hardware and they can't get new applications and so

forth and so on. And so closed source users are software slaves. Okay. Now open source is actually not much. Open source is great for developers because developers can get all this software that's already been written and tested and works and made with standards and stuff like that. That's wonderful. Saves them a lot of time and a lot of back in 1969 if you wanted to have functionality

you wrote it yourself. There weren't any databases. the word, you know, nothing like that. You had to write everything. But hey, you know, today's nice. We got all these big packages of software that we can use and it's great and it's tested more or less at least, you know, kind of and everything. And the developers, >> pardon me, >> works on my laptop. >> Works on my

laptop and developers know how to use it. Okay. More or less. But the end users still can't you still only get binaries in a lot of cases. They still can't change the software. They still can't fix the software. You know, the end users are still software slaves. And many of the detriments of closed source still exist. You can see generally how the software works, but not exactly

how the software works. So freedom software or some people call free software is truly the path. Free software allows the endusers to do all these things to fix the software to see how it works to upgrade the software when they want to and put as many users on as they use the software for any purpose Now I still remember the arguments back in the days of you

know developing the the four laws of free software freedom and people said well I don't want my software to be used by the military. I banks. I don't want my software to be used by the government. I don't want any my software to be used by the Catholics for God's sakes. Not the Catholics. So, you know, so when you drew the ven diagram of all the people

that couldn't use the software, you found out that it was nowhere. Nobody could use it for any reason. So, we obviously had to say you could use the software for any reason And we realize that you can't teach morality with technology. You have to teach morality by being moral. Okay? So when you have all this software and you can fix the things down to the last knit,

it now becomes a business decision. Maybe you don't have the expertise to fix the software, but you can hire somebody who does. This becomes your business decision about whether you live with the software the way it is or whether you hire somebody to fix it and maybe join in a community of people who will join together and fix that Now business people and I've talked to lots

of them. I talked to lots of them about free software. I'll tell you they hate freedom. They absolutely hate freedom. Government hates freedom too. And the reason they hate freedom is because with freedom comes choice. Okay? Freedom means that you can go and you can find another supplier. Freedom means oh god, you can't have freedom. What business people really like is control. If you go to talk

to a business person about using free software, don't even say free software. Just say you have control. You can now make the business decision about what you're going to do with your software. that they understand. So now we come back to sovereignty because that's what the name of this talk is, sovereignty. Okay? And I'll tell you, I love Donald Trump because well, so I used to love

Microsoft because of Vista. Vista was the best thing that Microsoft could ever do for Linux, except of course Windows 11. Windows 11 is taking the cake, okay? And even if you buy a new system with Windows 11 on it, right, it comes with it. They've got they they've got their AI so twisted into it that it slows down the system. And people have had the system with

Windows 10 on it, they put Windows 11 on it and it just slows down, right? And they they complain about it. So, and and and Lenovo, who is selling Windows 11 on their laptops, they're saying that they get one out of every laptop, eight laptops sold returned to them because the people complain that they're too slow and it's mostly because of Windows 11. Okay. So the word

you know Trump on the other hand has pissed off a lot of these companies so countries so much that they are now saying hey we don't want to use AWS we don't want to use you know Google stuff we don't want to use Microsoft because these companies are headquartered in the United States and if the US government goes to them with a subpoena and says either you

do this or you're going to jail well they have a choice. do that or go to jail or fight it off in the courts which cost lots of money and stuff like that. And so the rest of the world is now looking at sovereignty, okay, moving their data outside of the US, moving their their control of their software outside of the US. And I started thinking about

this. I said, well, you know, it's not just these foreign countries. People need sovereignty. People need to have control of their data. And it's not just their data. They need to have control of their software. networking. They need to have control of their technology, the design and the manufacturing supply chain. How many of you remember during COVID all of a sudden you couldn't buy a new car.

You couldn't buy a new car because all those tiny little microchips were tied up in some boat in some country and stuff like that. we couldn't get a hold of them and that they couldn't get a new car and then you couldn't get a used car because all the people that wanted used cars were now buying used cars and everything was drying up. This is a supply

chain problem and we're going to start seeing a supply chain problem again. We're seeing a slight one because of tariffs, but we're going to see it again because of the war that's happening that's cutting off the oil in the straight. Okay, we're going to see this happen again. users want to have control. Now, I think you get the idea here and you think you get the idea

of sovereignty. But there's a problem that's coming up and I want to warn all of you about it. That problem is called quantum computing. And it used to be that we would say, oh, you know, why is quantum computing bad? is because all the stuff that we have spent all the years encrypting and we have spent all the years building VPNs that is about to be undone

by quantum computers that are going to be able to come in and decrypt that stuff that we thought was going to take a 100,000 years to decrypt is going to be able to decrypt it in 15 minutes. And so if you say well you know this is still a long time away. We'll get to that in a moment. The thing is that what these people are doing

is they're capturing your data now and they're selling it on the data market. They're holding it ready for that quantum computer to be ready to decrypt it. Okay. So, we need to start right now in recent computing safe or at least resistant algorithms that NIST has made public that people are starting to to use. Red Hat has started to use it, you know, you know, they make

it available and stuff, but you have to use it. You have to be aware of this because it's coming. Just like Y2K, you know, came, you know, we managed to do enough stuff to fend it off, but we need to start now. And even now may be too late. Okay? And that's because when we looked at the next month next month it was two years out. Okay.

And now Brazil is buying a commercial quantum computer from a Chinese firm. Okay. So now it's turning into a commercial thing that people it's still a small one still not very powerful. Still maybe for education but it's coming. Okay. And it and like I said, it may be too late because they're harvesting the data now and they'll decrypt it tomorrow. And things like, you know, your bank

account numbers, you know, every time we turn around, we see another bank that's been broken into and somebody's captured the data and stuff. There's a value for that and people are selling it. This includes in So, you're sitting there, you're safe. I'm using my VPN. I'm safe from people spying on me. Okay, you're in an airport or something. But let me tell you, when quantum computing is

there, while you're in your encrypted system, while you're you're using your VPN, they could decrypt it, find out the keys, and break into it, and you wouldn't even know it had happened. So now people are beginning to say, what cloud should we use? And the question you have to ask yourself is, what laws apply to it? Now, I do a lot of work in Brazil. I'm married

to a Brazilian. I've been doing a lot of work years and they have their own laws about encryption and data and privacy and stuff. And normally that would be good enough for them. They're laws of their country. They vote on it. They can change it. They can have it. But they're also under the laws of the United States because of the Patriot Act and things. So I

as a US citizen am protected by our constitution from people spying on me. But under the Patriot Act, if I talk to somebody in Brazil, then they can they can come in and watch our conversation. And that I object to. And we need to be able to say what supply chain are they using for these clouds? Are they going to be able to keep up with this?

Now recently we had a whole thing about lack of RAM and lack of storage because of AI. Okay. Well, the thing is there's a storage people and the RAM people say we're sold out of RAM and storage for the next two years. But what happens if the AI bubble crashes? What happens to all those contracts they have for all that data and stuff? Who's going to be

hurt by that? We need to think about our supply chains. And if if we really need that much more RAM and that much more storage, we need to start building those factories now so that in two to three years we'll be able to have the supply we need. Now what chip do they what chipset do they use and how do you test that chip? Now most Intel

chips as an example are called complicated instruction set computers that for every machine language instruction you have hundreds sometimes tens sometimes hundreds of these little micro instructions which are being executed inside the chip. You can't really test you can't really see what that micro code is doing. And you could find that there would be some very large, very powerful entity, let's put it that way, who might

grab some board that you've done, pull it apart, chip that they have made into that board, and then packaged it up, and you would never be able to tell what was opened in the first place. Now, if you think this is impossible to do, let me assure you it is not. Okay? And I have actually ran into people who can do this So, and and if you

might remember a couple months back, maybe a year back, there were these pagers that blew up in Iran because a certain government had gone in and changed the pacer pagers and programmed them and shipped them out and had them purchase the stuff and they hit the button and boom, all these pagers that they knew who had bought them blew up and killed people. This is the type

of level of sophistication that we're going against. Now you might say, you know, I as a person, as a person on the street, I can't really have the opportunity to know this type of work or to test this type of work. However, your government should be doing that. Your government should know where the fabrication plant is and where is the PCB board manufactured and what is the

boot code. Is your boot code trusted? And to do that you need to have open-source for all of that stuff. So today we have two things that are fun fundamental in this. We have freedom software where you get all the source code and we have risk 5 architecture which is a so it's a reduced instruction set computer that you can actually look at the mask and see

how it is done how it's manufactured and then because each instruction is typically one machine cycle you can actually time those machine cycles to see whether that chip was fixed fooled with or not. Okay, it's not perfect, but it's a lot better than that. So, what we need are sovereign clouds. Clouds that are put together with security first. It is the first thing that you you think

about. Now, I'm going to tell you why a lot of the modern-day clouds are not They're not security first because you have to log into them. You are on your whatever system and you have to give your name and your password to go across the internet and log into that system. Well, if you can log into that system with your login name and your password, somebody else

can log into that system with your login and password. In fact, they can find a a a level of attack on that system to go in even without knowing your login and password. Okay? So, this is not security first. This is ease of use first. This is availability first. It is not security first. clouds where people can't interfere with your path from yourself to where the cloud

is. This means local data, local cloud, local systems. A lot of cloudfacturers are doing that. They're they're setting up local data centers, you know, around the world. However, they're still all controlled really by the companies in the United States. And you need authentication that is both open and secure. Now I consult with a company called Q Sentinel and they decided to put together a cloud service that

is security first and is sovereignty first. In fact they want to have they tell people they give them so sovereignty. So they've set up a data center in Switzerland where they're using mainframe computers and they put their software on top of Red Hat Linux running on top of VM IBM's VM running on top of uh of uh M uh IBM Z series computers that gives a very

very good security model but the interesting part of their Security is you don't log into their server system. Their server system logs into you. So you have devices which capable of creating a quantum proof VPN and the server administrator knows what those where those devices are and creates that V that QPN to the device. There is no attack vector. This thing is invisible on the internet. You

can't log into it. You have to use that device which is known to be able to talk to the cloud and therefore you log into your device. That's it. You log into your device and then the device is connected to the cloud. And this is a service that they're selling right away. And you know it's for big companies, it's for you know it's for governments. is for

people who really really really need that level of security. However, not all of us are there. We want to have a freedombox.org. This was started by a guy named Evan Mowglin um back around 12 or 13 years ago. Eban is a lawyer at Columbia University. He is a security and privacy person and he he was very worried about you know cloud stuff. He said I want to

have a really secure cloud that people could plug into their wall. It uses only like five or six watts of electrical power and I want it to be easy enough that people would be able to set it up and use it. Is it mom and pop easy? No. But is it a lot easier than setting up all this stuff that you need for a secure system? Yes.

Okay. And about two years ago, they came out with version 1.0 and started to publicize it and people started to look at it and it is it's written to run on single board computers or recycle computers. Okay. It runs on almost anything. And the software for it is included in Debian 10 and it may be included in Fedora. I haven't looked. I don't know. So you can

set up your own server and then you can also do a quick a neat backup thing where you set up half your disc space for somebody else who has a freedom box and they back up their freedom box to your freedom box and then you back up your data to their freedom box. Okay? So if your house catches on fire or stuff like that, you have a

backup that exists some other place. If there's a flood or a famine or somebody breaks in your house and steals it, you got your backup some other place. Now, this also supports the Fedverse.party software. So, you can set up a little community type of server that people can share and share their most private things. If you want to send all of your cat pictures to Microsoft Cloud,

go You know, just inundate them with cat pictures. I'm sure they'll love it. But if you want to have your things like your credit card numbers and stuff like that, you want to put them in this really secure little server. Now, recently I started looking at multiffactor authentication. And you know, when you when you have that's one level of authentication. And fortunately, over the years, we've gotten

a little bit better. We no longer use passwords. We use past phrases. Okay? my passphrase. I guarantee you that you will go to the dawn of time and never be able to guess it. Okay? Never be able to break into my system using my passphrase. Okay? Hm. But in any case, there's lots of other types of authentication. So, for those of you that use a cell phone,

maybe you have a fingerprint on that. Uh my cell phone also allows facial recognition, although I don't I wouldn't torture anything with that. And and I'm not a big fan of facial recognition either because of the large amount of hair that I have on my face. But you can also have different types of like if somebody's trying to log into your system and they're they're not in

your country or in your state or in your house and it's not a certain time of the day or day of the week, then you don't let them log in. Okay? And these are all these different types of authentication that you can build into your system. Now I started looking at hardware keys and there's various different types of hardware keys out there. There's um uh UB key

as one and these are little hardware keys you can plug into your system that allow you to log in. You can also set it up so unless a hard keyware key is present you can't do a sudo onto your system. You can't switch users onto your system. You can say, "Hey, if the hardware key is present, I will de- enrypt my file system." Okay? And so all

these different things that you can program into your system and it follow they follow a standard called 502. And you can even say, "Oh, I'm at a I'm at a uh a system in the in the local library. It's a Microsoft system and I want to use it. I mean, I don't have a login and password on the on the on the system in the library, but

I want to log into my Google account. And so, you set it up so that it will look for this on the system in the library and allow you to log into Google. Without it, you can't log in to Google. Okay? So, it can be used by lots of different services in various different types of clouds or it can be used to protect your system. This works

on Linux, BSD, Windows, Android, iOS, all the different major systems in So this Oh, I I I was I was talking I was looking at the wrong side. So if you get a hardware key like this, whether it be a UB key or this particular one is solo key. If you get that, then you need two of them. You need one that you use for every day.

And I normally keep this one on my on my keychain. Uh, and you give one to your mother or you keep it in a safety deposit box because if you have one stolen and you lose the other one, you're cooked. Okay? There's nobody who can help you. So you need to have one in a safe place so that if you do lose your primary key, you can

capture the other one and then remove the one that was stolen or damaged or anything and add a new one in for authentication and then you give the new one to your mother or your safe deposit Now, in the best of all possible worlds, this should also be an open design because you want to be able to have control over it. Okay? And so, I've been working

with solar keys.com, which makes an opensource key. And the firmware, the source code it uses to put in your system, it's all open. They have they publish how their PCB is made. They have their Gerber files. They have their bill of materials. And so if you're in a country, this maybe not the United States, you say, "I want to manufacture this in my country, you can." Okay?

And it's 502 compliant, so it meets the standards and everything. You can use it through SE Linux for Red Hat or if you want to use App Armor is supported by PAM and App Armor. And so, you know, it's very, very efficient, very good. and LPI is kind of helping them market it because it is free and open source. It is open and if you go to

our booth in the exhibition hall or you can take a picture of that right there, you'll get a 23% discount on everything in their in their um in their catalog. And there's no limit on how many of them you're going to get. So, if you want to buy a thousand of them for your for your company, then you'll get 23% discount, which is about 10 bucks for

each one. So, you save $1,000 or $10,000 on it. And so, that's basically my talk. And if you have any questions or, you know, anything like that, at our booth, we have the equivalent of that little sign Oops. I didn't want to do that. We have the equivalent of that little sign. So, if you didn't get a chance to take that picture, you can come by our

booth and take a picture of the sign there. And now what I'm going to do is I'm going to go back up through all my slides and remind people that the old farts bough is in room 209 between 1900 and 2000. And um I tried to get beer there, but we just didn't work. It just didn't work out. And for Windows users, that is from 7:00 p.m.

to 8:00 p.m. And so with that, I will ask for any questions. And if you have a question, um, please say it slowly and I'll repeat it or we can maybe get the microphone sent around to you so you can ask it and get your voice recorded for posterity. Hi, thank you for the talk. Can you hear >> Yes. Can I hear you fine? >> Um, you

know, I don't know why I wrote down this. Did you talk about there's a some kind of timing time on the chip on one of your slides? >> I'm sorry. What? >> Uh, time. There's for the hardware chip. There's a way you can time the chip or >> Yes. You can select like >> what? Wait a bit. You need you need a microphone. >> Were you talking

about what what he was saying about if someone's trying to log in as you, but it's the wrong time that you would be blocked. >> Oh, the risk. Okay. >> I I know I know I know the question. I will answer it. Okay. I'll answer it. So when I was working for Digital Equipment Corporation, we had a program called Axe. And Axe would go through and execute

every instruction in the instruction base, every addressing mode, every data type for the entire vax instruction set, which was a complicated running at a million instructions a second. It took an entire month to run that program one time. Okay. Now, if you reduce the number of instructions and you reduce the number of addressing modes and you reduce the number of of data types, you could run the

same program in 30 minutes. Okay? Because it's kind of exponential as as it goes. Now, this is just to test each instruction for the proper length. Now let's say you know um in the IBM instruction set as an example the two instructions that were the the longest running instructions were um changing a uh an epsidic number like character epsidic character to a binary coded decimal. Then the

next instruction was changing the binary coded decimal to a binary number. These are very long and and instructions and you could time how long those instructions were running. So if you change the microode for those instructions, it would probably change the timing for that and you would be able to know that somebody had tampered with that instruction because risk instructions typically only have one clock cycle per

instruction. They don't have any microode. You can't really test that, right? It doesn't. It's really not possible. So, people would typically not change a risk instruction to do that. Okay. >> Thanks. >> You're welcome. I know all these weird things. Okay. >> Test testim. Yeah. >> Oh, hello. Um, you said about those um the solo keys, right? Well, interestingly, I was one of the Kickstarter backers for

those um when they were doing the Kickstarter. >> Mh. >> And unfortunately, COVID happened and there was a big delay >> I got the keys probably a year or two >> Mhm. >> And I had the one one of the very limited edition purple keys that they actually had, very limited purple ones. >> Mh. And um for and then for some reason right around 2023 and on

GitHub people are questioning is the project dead? There are no firmware updates since like 2022 when I last updated. >> So I was surprised that you were doing this and it's like I was kind of afraid >> you know solo keys was dead and >> I've actually moved on went to UB Keys. >> Okay. So let me tell you what happened. >> Do you know what happened?

Yes, I do know what happened. So, I was working for the Centinel project and I was also one of the Kickstarter people on that and there were a lot of Kickstarter people because they had certain problems in doing their manufacturing. One of the problems was each one of these keys has a little indentation that they pour epoxy into. So you can't change the components on the key

without people seeing that it's been tampered with. And getting that machine to be able to do that was one of the things that held up the production because they were expecting that it was going to happen. You know, the Kickstarter project would last a year. It lasted only two years. And a lot of people asked for their money back and they gave them the money back. But

they eventually did manufacture the the chips. And that's one of the ones you got. I got back then. But then COVID hit and disrupted their supply chain and disrupted their sales of their chips, their keys. And so they they they didn't they couldn't go on. They just stopped. And yeah, it was kind of dead. But now the same people have started it back up again and they're

determined that they're going to make these keys and make these keys go forward. And so COVID is over. They've they've redesigned the key because there was a a mistake in the uh NFC chip. They've gotten the new chip. They've redesigned the print circuit board and they're moving forward. So, I got involved just at the time they were making this decision to go forward and now they're going

to start manufacturing them again and selling them again and upgrading their their website again and upgrading the software and stuff like that for everybody that had had the old keys and now have the buying the new So, I have to look at the future. I can't look at the past. But I have talked with these people. I believe that they they have their hearts in the right

place that they want to do continue doing free and open source stuff can continue publishing their circuit diagrams and their software up on the internet. And so I say, "Yeah, they're normally $46 a piece, you know, single single quantity, but you can get uh 23% off if you buy them through this through this thing." So I am also working with the University of South Paulo. I'm working

with the Freedom Box people. Freedom Box is kind of in the same boat. They they're a great thing, but they haven't had a lot of publicity and so the number sales that they have are small and I'm trying to get them to have larger number of sales to generate more money that they can have more thing in the in the project and that's what we have to

do with free software. I will point out that one of the problems of free software on the desktop is that for the c a c certain number of years we would have 2 to 3% of the desktop sales and so we didn't get things like games and stuff. We didn't get a lot of the applications that we really wanted to have ported to free software because there

wasn't enough of a market. And so to fix that problem, you have to create the market. And right now, we're selling between 11 and and 15% of the desktop sales. And that's going to start to attract a lot of the companies that have imported to to desktop free software to do that work. And so what I'm hoping and there's, you know, I've been in the business for

50 years. I know that there is a hump that you go over and when you reach that hump all of a sudden lots of doors open up and we need to be able to say, "Yeah, we're going to push this." Every single one of us has to go out there and push this. I know that I'm preaching to the choir here in this room, but there are

some people this may be their first scale that you've come that they've come to. So we have to tell you the warts as well as the as the benefits and you know you need to go out to your to your educators your government and you have to say are you going to be slaves or you're going to be masters of your fate and you know there's a

lot of stuff that's happening right now that is an opportunity for us to have more Linux And maybe you're not in the United States, maybe you're in Brazil, maybe you're in Ghana, and you have an opportunity now to work with free software and to build an economy in your own country. That to me, the far-left Bernie boy is the way that we really get world peace. when

everybody feels that they can have an opportunity to make money, to make a living, and to have control of their own destiny. Okay, we have about 20 minutes. If we can still take more questions. >> Um, sorry to >> sorry. Sorry to dwell on the solo keys more, but um Nitro Key makes a similar authentication key and I believe that they're open source. So is that uh

I should say yeah I don't know open source free I know I don't know if it's free and open source but anyways um do you know any differentiators between the solo key and the nitro key? Uh no I haven't taken a huge look at all I know I do know there are other keys that are open source but again you know it's it's to what is open

source and what is free and I know that solo key but talking with the management stuff like that they feel I'm not saying that you should go out and buy solo key I'm saying that we're selling it here with a 23% discount you can look at it yourself you can make your own decision about which way you want to go but what I am telling do is

you really need to think about multiffactor authentication and really you know make that part of your security platform. Okay. So you you know you I mentioned Ubi key solo key you mentioned nitro key there's all different ones but the the other thing about it is they should be following a standard like 502 so that if for some reason one of these companies goes away your MF your

your multiffactor authentication is still valid. Okay. So, thinking of the cloud, my feeling is that I don't want anything that I depend on to depend on something that has to go to the cloud. If I want a self-driving car, I want the car to do all of the work internally. no external communication at all. If that works, it's no different than if I'm driving the call car

because I have no external communication while I'm doing that. And so I have the same amount of control. But if you go to the cloud or you go to any kind of communication, you do not have that autonomy. >> Well, that is true. But back in the day when I started programming on computers, you had a computer in the room and networking was carrying a box of

cards down the hallway to your friend. Computer security was locking the door at night and graphics was printing out as art of Snoopy lying on top of his dog So after that, we had some type of communication. In the early days of Unix, it was Unix to Unix copy UUCP. And that was from node to node. Okay? And that's what we used. But then after a while,

people started saying, well, we should have servers of workstations. And you hooked up your workstation to your servers so you could communicate your data faster and better. And you could also do larger problems that you wouldn't expend on any one of your given desktop systems. Now somewhere along the line they changed the word server to cloud. And all the people who used to do client server you

stuff said well the cloud is just another server. And so the little uh cloud I showed you, cloud server. I showed you the freedom box is a cloud. It provides you cloud services. You know, it was interesting that for a while I looked at Freedom Box and I said the software in there is really clunky. It doesn't have a good user interface and stuff like that. And

then all of a sudden I realized it's not supposed to be a user interface. is supposed to be a cloud services thing. The user interface is still on your iPhone or your Android phone or your or your Linux desktop. That's supposed to supply the user interface. And all the cloud is supposed to do is transfer blocks of data to you so you can share those blocks of

data with other people. Now, the real question is who controls that cloud? Is it you or is it somebody you trust or is it somebody that's far away? And who can turn off your cloud? And where do you get your updates from your cloud for your cloud? And that's what you have to determine, you know, but I think there's been uh evidence of people or or clouds

that people break into them, steal a lot of data. you never even hear about it, right? And I think that all the pe all the instances of of of clouds being broken into for all the ones we've heard about, there may be a lot more of them that we don't hear about that are hidden from us. So these are all things you need to think about. And

you know I I understand your feelings sir but you know I go around I tell people that sovereignty of computing is so important because if I click my fingers like this and computers were to disappear from the face of the earth then these lights would go off airplanes would fall from the sky. Cars would refuse to run. We're stuck with it. We can no longer compute the

needs of the world with a slide rule. In fact, most of the people I know don't even know how to use a slide rule. I do because I'm an old fart. So, did I did I answer your question, sir? Well, I hope you don't use email then. >> I just have a question. How do you think the future of AI looks like and how can we steer

it towards individual sovereignty? >> First of all, I was using expert systems back in 1986. In fact, I wrote an expert system to be an expert systems administrator. An expert system at that time, the one I was using was using a language called Eclipse developed by NASA, which was And what happened was you had rules and you had uh you had rules and you had data and

you applied the rules to data to see if the data was conforming to your rules. So what type of data were we talking to? Sorry, not data but facts. You had facts and you had rules. What type of facts were you talking about? Well, it was the permissions on a file that was a fact. It was the path length of of the of the path was a

fact. It was the count of how many bytes were in that file was a fact. Okay, these were all facts. And if they differed from what the rules said they were supposed to be, then maybe your system was compromised. And so you developed rules to change the system back to what it should be. and clips was so fast that I could go down through a Unix system,

the TCP IP and UUCP uh systems of a system in about a third of a second and that was running at a million instructions a second on a VEX. So that was an expert system and we had lots of expert systems. doctors were using expert systems. And when you went to a doctor, they looked at temperature, your eyes and muc. Then they ran into their back room

and they opened up this big and said, "If you have these symptoms, you probably have this or this and you need to test for this or this to see what it is." That's neone of these show any real intelligence because intelligence makes new things out of nothing. Newton was intelligent. He wanted to he wanted to describe gravity and he couldn't do it because he didn't have calculus.

So he had to invent calculus. Right? That's intelligence. Okay? And I think that what we're seeing today as AI is actually a bunch of expert systems which are creating their own rule bases, but that isn't necessarily intelligence. So I think we still have a long way to go. And as a person who also takes a look at the human mind and the number of neurons in the

human mind and the number of synapses in the human mind and connections in the human mind and in fact if you take these synapses and line them end to end it'll go around the earth a couple times we're still a long way away from something that can really create intelligence and we don't even know how we think. Okay. So, a lot of these people, I think they

developed the term, you know, artificial intelligence, and they're selling it to make a lot of And the question is to me, when do people realize what the limitations of this is, and how badly is the market going to crash? That's a long- winded explanation. Uh, you're talking to the wrong person if you're looking for somebody who really likes AI. Now I'll admit I've seen AI do some

pretty amazing things and it helps to sort through you know might be closer to facts but the people that are using it to write code the problem I have with that is if you ask it to write a piece of code today and then two days later you ask it to write the same piece of code it's going to be different code and so how do you

maintain that? How do you fix that tiny little thing that's wrong and be able to test that tiny little change that you need to test? That's a problem they haven't solved And you know, so I'm not a I'm not a big fan really of artificial intelligence. I think that more and more people are beginning to question the artificial intelligence model, but we'll see. I'm 75. I probably

don't have that many more years to do, but my husband, he's only 31. He will. Any other questions? >> Great presentation. Can you talk a little bit more about the uh the quantum computing story that you gave us? uh the the the risk of it. If you're saying it's already been collecting, how bad is it for the for those out there that did not protect and what

do we do going forward to prepare for that? The problem is we don't know because there's been so many people who have broken in and simply stolen the data and the data was stolen. They say, "Well, it's encrypted. It's protected. It's stuff." I mean, when I started using RSA, I had a key length of 512 bytes. Then it went to 1,024, then 148, then 9,000, I'm sorry,

not bytes, but bits. And you know, I kept increasing it because computers kept getting faster and faster and able to break it in a shorter and shorter period of time. Quantum computing breaks that completely. Okay? Because it's made up of two prime numbers. And the problem is the the quantum computing can test each one of those combin possible combinations of prime numbers so fast that it's not

hundreds of thousands of years anymore. It's 15 minutes now. It takes a certain power of quantum computing to do that right. And the thing is with quantum computing it generates millions and millions of answers but only a certain number of them are correct. And so what you do is you filter out all those incorrect numbers. And that's what they're working on right now. They're working on better

algorithms to do that. They're working on having more what they call Qbits in a in a processor to do that. There's lots of problems, but they keep working to solving them. And all I'm saying is we know we do have algorithms that quantum computers are not going to be able to break, at least not the way that they could break the ones we're using. So why not

start using them? Okay, you know, think about it. Start making start taking your data and encrypting it in a way that will be protecting you from when these quantum computers come out in force. Don't wait for the year 2000. Anybody else? Thank you all for coming. >> Don't forget the old buffs, the old >> So, what's the cut off for old for old parts? >> Well, I'm

gonna I'm gonna I'm gonna start I'm going to start I think I'm just a guide on this. I don't want to be the center of this thing, right? >> So, I'm just a I'm just the the guide. And uh I'm going to say okay how many of you started programming at this time and you know and >> but it's not it's you know the old farts thing

is not your age or how how long ago you started a lot of it's influenced by your experience and stuff like that. So you would definitely be one person who could pipe up and say no. But what what inspired it? What inspired it is I've been listening to to discussions on Facebook about things like the the FSF and the GNU project and how it doesn't have any

relevance anymore. I say you're right? You know, if it had that I mean people say, well, what about BSD? Well, wait a minute. BSD was still being sued by USL in September of 1994. >> If I had waited for that suit to go away, NT would have won. I I couldn't wait. I didn't know when that was going to go away or if that was going to

be solved. So, you know, Oh, hello. Yep. All right. Sweet. Hello. Hello. Hello. Sounds like Yeah. Yeah. I'm amplified. Hello. test. Oh, I have a cell phone next to my next to the headphone wire. Like, are you getting like interference or something? That was I'm presuming >> it's not critical. It just needs to the machines go kachchunk kachchunk kach chunk, you know. Yeah. Oh, I'm now hear

>> Try plugging in. All right. >> It's plugged in. >> Is it making that sound? >> Thank you guys. All right, we are at time. Uh, thank you everybody for showing up to the last talk of Saturday. I mean, what a big Saturday if you went. That's right. Yeah. Um, so, uh, yeah. So, I really appreciate it because you could be having dinner right now. Um, but

you maybe you had a late lunch, so that's cool. Or you're you have or are you have a snack, that's great. Um, I my name is Kyle Ranken. Um, I have been involved in Linux for a long time. Sort of cut my teeth as assisted and kind of pivoted over to security halfway through my career. I kind of have a leg in both, I guess you would

say. um also an author so I wrote a number of books about Linux over the years and was a columnist for about 10 years for Linux journal magazine. Um also I've been at scale a lot so I found out I thought I realized earlier today that this is my 15th scale that I've been to and the 15th one I presented at. Uh some of my first presentations

were like back at 8x um which I had to do a little bit of math uh to figure that out which is a great segue um into this talk which is about talking about the UI lessons that we can get from antique calculators. Um all right so let's jump right into So uh brief introduction uh it's really easy for us to take uh UI like user interfaces

for granted. So uh a lot of people when they talk about UI they will talk about things being intuitive. Uh but intuitive is really a relative term. What does that actually mean? It probably means it's similar to something you already know. um not necess there's very few things that are actually intuitive. So um like for instance let's take the calculator. So this is a user interface for

a calculator. For many of you you might consider this the user interface for a calculator. If if you were to be presented with this most of us are of an age in which in school you were given one of these things and at some point in in math courses and used it to do some calculations after you learned how to do it on paper maybe. Um, but

you would you use this is and if I gave you this kind of a user interface or something similarly shaped, you'd be able to and it gave you some math problems, some basic arithmetic, maybe even some multiplication, you'd be able to crank it out on this, no problem. Right? So, one I have like 5,000 hobbies. Uh, but one of my hobbies is that I collect vintage calculators.

All right. um things going back for the last you know for mostly consumer consumer models. So I don't have like anything from like the like early 1800s or anything. Um but I started it started down the road of actually getting something like this. This is like a handheld pocket calculator that would fit in your shirt pocket. Um after this people can come up and try this out

if you want. I got one of these and I was just fascinated by how inside here there's no computer, right? But it's doing math. Uh, and it's doing math purely mechanically. And just that notion fascinated me and I wanted to learn more. And then I went down the rabbit hole that I always do. And now um my office is kind of an antique calculator museum. So uh

I have between 30 and 40 uh different uh antique calculators in my collection of different shapes and sizes. Um, and when you uh collect vintage calculators, what you will discover is that the user interface of a vintage calculator varies widely across models. And I mean widely. And I don't mean, oh, well, the plus sign's a little bit bigger. It's in a different location. I mean, they're completely

different. In fact, you and I, we're at scale. I like very intelligent audience. Um, I believe that many of you, if not most of you, would struggle if I were to give you um, any of my calculators in my collection, probably even including the one right here in my hand, um, and asked you to do a basic math problem without instructions. If I just handed it to

you and and said do 12 x 12 or even maybe even 12 12, uh, you might struggle to do it without instructions. All right? And it's not, this isn't because of any personal failing of yours. I found it the same way when I started getting these things. I I would pick up one and like how do I use this one? It's completely different from all the others.

And this is because mechanical calculators, especially the desktop models are incredibly complicated. They work via very intricate complicated sets of mechanisms internally that perform calculations and do things like, oh, if I add a number to nine, I need to carry the one to the next register and all that kinds of stuff. It does it all mechanically things that we maybe take for granted unless you're doing it

on pen and paper. Um so very complicated mechanism. When people were design were building and designing these calculators they weren't focused on the user interface except that they were really they were focused on getting the machine itself to work. Can I create a machine that uses gears and and mechanical mechanisms to perform math? All right. The UI was essentially how the engineer exposed the inner workings of

this machine to the user. And that's it. It was like, and you'll see examples here, but that's pretty much what the UI was. Um, basically this is similar to how most Linux apps are developed, right? a developer writes some software internally. They've structured their software with functions that do things. They inter they in understand the internal data structure and when it gets around time to expose the

UI to a user they who deeply understand the software simply expose those functions. If it's a command line tool every function is now a command line argument and the arguments that went to that function are probably arguments to the command line argument etc. Right? There's not really a whole lot of thought given to how would a someone who's not me writing the software use it. So this

talk we're going to do a dive through the history of antique calculators. Um I'm going to give a number of different examples. Each of these have a distinct UI. Um these are all well most of these are ones in my collection. I'll call out the ones that I don't yet have in my collection. Uh we're going to start back in 1874 with the original um and go

all the way through, you know, comtoers, which are fantastic. Oh wow, it's really loud now. Sorry about that. um all the way through to the digital age and then at the end kind of like wrap it up with like make it about Linux you know talk about Linux a little bit and but I think you'll find it pretty obvious with the setup that I did each phase

of this uh you'll probably make your own parallels uh to Linux UIs you've used all right let's4 looks like this all right um this is probably one of the first consumer calculators so you can go way back way before 1974 and talk about mechanical computation. This is this goes back along the Charles Babage etc. Like you can go way back. Um and there are a lot of

this isn't like the first example of a mechanical calculator but this is one of the early consumer examples where you would have someone I mean there's a couple of other examples but this is one really took off and one that you still actually would see for many many decades afterwards. Um this is known as a pinwheel calculator or a pin wheel mechanism. That's because if you were

to look at the look at it from the inside and take it apart, the mechanisms that allow it to perform math look like a pin wheel and they rotate and there are mechanisms internally to carry ones and do all of that great stuff. This calculator um can perform the four main types of arithmetic. So adding, subtracting, multiplying, and dividing uh which you may might assume is a

core competency of a calculator, but you're you're taking them for granted. Um but this one can do all four and because it can do all four it can do many of the other things that if you are clever you can use those four main arithmet arithmetic functions to perform. Um to set numbers on this, you'll see at the very top there's a bunch of little like and

there's these little So to use this calculator, you would set your first number, your first digits by sliding the slider down to the appropriate digit one row at a time till you have your numbers. Then you would rotate there's next to the right is a is a handle. You would rotate that clockwise if you want to add and you'd rotate it counterclockwise if you want to subtract.

All right. If you want to multiply and divide, you do it kind of like you would do on pen and paper where you take uh for example uh if you want to multiply 12 x 12 on a piece of paper, you would start with the two and you say 12. Basically, you're calculating 12* 2. You write that down. You shift to the left and then you do

um 1* 12 essentially and you add it together and you get 144. Okay, so that's how you would do it here too. You would do 12 and then you would shift and you would do it twice and then you'd shift and then you do it again. So that's how you do multiplying. Dividing is the opposite. Um you shift all the way uh to the left. Um you

would put your your um dividend in to the calculator by adding it to the register. Then you in you reset set your divisor and then you start subtracting like crazy until you hear it ding. Um there's an internal bell that when it underflows uh maybe I marked this. Yeah, maybe I did. Yeah, I did. But doesn't matter. You would crank it and subtract until you underflow, which

means you've subtracted so much it goes into the negatives. That triggers an internal mechanism that rings a bell. Um when you hear that bell, you now know, whoops, I underflowed. need to add one more and shift. Then you subtract, subtract, subtract. Ding. Add, shift, repeat until you run out of digits. Um, and whatever's left in the register is your remainder. And whatever's on the actual main register

is your answer, your quotient. So, um, when you're done and you want to get you want to clear instead of pressing the AC button on your calculator, um, you see a couple of cranks at the bottom there that you turn them and each one clears a different register. And I said this already, but there's an internal bell, like a physical bell that aids the vision. So, what's

this like in practice? So, I don't have an original Odner from 1874. This is the oldest model in my collection. This is a Brunva, which is a clone. Um, a different brand, a different company made a uh calculator along the same lines using the same uh design. Uh, so this is from the around the 1890s. And I'm going to demonstrate now um how to multiply 12 x

12 using a princeva setting the registers. All right. Now we shift. There we go. 144. All right. 12* 12. All right, so let's scoot ahead not that far later and we have the comtoter. The comtoter. I'm very excited about comtometers. I wasn't originally excited about comtometers. They look like this. Um, you see this, you're like, whatever, man. Like that's that's that's not nearly as cool as that

other device that has these things flipping around or whatever. The comtoter is sweet. I no arguments. You're raise your hand. It's not going to be answered. Comptoter is awesome. Um, but it doesn't seem like it seems pretty plain. But this is a speed demon. This is probably the fastest mechanical calculator ever made. Um, and I would argue in the right hands possibly even faster for some operations

than some of the electrical mechanical calculators like some I'm going to show you later for certain types of operations. Um relatively simple design. Um there's some basic inputs uh that you have. You see these buttons here. There's not a whole lot going on, right? Um you can there's is mode switching. So you're switching modes between adding and subtracting. Uh there's a really steep learning curve uh for

this calculator. Um steeper I think than some others. Um particular there's some training manuals. Uh there's a lot of training manual. There's some vocational courses. In fact, I have like a little award pin that you would get in high school if you took a comtometer training course so that you would have on the job training so that you would be employable. Um, you know, sort of like

the CS courses you get in high school now. Um, you would get a little pin like I can get a job uh operating the comtoter uh when I get out of high school. Basically, this was the VI of mechanical steep learning curve, but once you learn how to use a boy, faster than anything else. All right. >> Yeah. And that Yeah, Bill said correctly that it explains

a lot by why I like it. And yeah, that's true. Um um it can perform the the four main types of arithmetic. You like, well, how in the world can you do that? You can. Um there are manuals that talk about how to perform square roots on the comtoter. it's deceptively simple to add numbers. You press the button down that corresponds to the number you want to

add. Um, that's it. Press the number down, it adds it to the register. Boom. To subtract though, um, you hold the subtract lever. Uh, you see these little metal bars down near the bottom above the numbers. Um, some of those are flat, so you can mark a decimal place or you can mark like a thousandth or whatever. Um, but the ones above that, the little shiny metal

ones are levers. And you up with your thumb. And as you do that and press down the numbers that you want to subtract, not the on the big numbers. If you squint these little teeny t ones complement and the thing about the ones complement is that is how you subtract through addition. All right? So if you want to subtract a number using addition you add the on's

complement and don't carry the one. If you do that then you end up with uh the subtraction. So the way works because it can only add all this machine does is you push buttons and it adds when you pull that lever it um just stops it from carrying the one. That's all it does. But you have to know um what the ones complement is or more specifically

you look at those numbers, those small numbers and you press those down. There's also a little to the side. All that's for is to clear the register. You're ready to clear. You throw the lever forward, pulls, and it resets everything to zero. And that's pretty much it. All right. trained operators who were also in their day known as computers because they perform computations um operated without looking.

All right. So you see an ad here. This is an ad for how comtoer was they chosen and this is one of the ones I have in my uh collection hanging on my wall because I'm a big nerd. Uh so this one is advertising for Zenith. But this room is Zenith's computer room. This is the room where all the computers were working. Um they as you can

see if you squint um they're all looking they're not looking at the comtoter. All right they are looking at a ledger with figures and they are touch typing the comtoter. All right professional comtometer operators always operated without looking. They because they had to focus on the ledger. They had to keep their finger on the figure that they were that they were uh adding or subtracting or whatever.

They you may notice, well that's weird. Like these old calculators, and this one in particular was the first one that we showed that I showed you that has like all the numbers, right? You're like, that's weird. I'm used to kind of like just having like zero through zero through nine, right? Um, but this used the full keyboard with corded key presses. They wouldn't if you they wanted

to add 123, they didn't go one, two, three. They went one, two, three all together. So they would cord all the numbers. And if they had to do multi-digits bigger than that, they would do it usually about three at a time. and then they would shift. And again, without looking, they would shift. Um, the other thing they did for extra speed was they stayed on home row.

I don't know if you noticed, but it's a pretty big long row of numbers. They didn't get above five. They would stay um the number seven, they would enter three and four. They would never enter seven because it was faster to hit three and four and stay on home row. It's the VI of calculators. I mean, come on. So later models um added error detection. So here's

the thing. Um the thing about a a device that automatically adds when you press it does if you fully depress the key all the way down. But what if you what happens if you only have to press one of the keys? Say maybe you're pressing three keys at once, but one of them doesn't go all the way down. Um the comtoter may not register that addition. Well,

then your calculations are off. Later on, added error correction into them, including the one that I showed the picture of before, it has error correction so that a press locks that register. So, the next time they go to press a digit that digit down anymore and they're like, "Oh, this whole row I need to investigate um and figure out and and clear the computation." Made in 1887.

These continued into production until the 1970s. This is how prevalent because until the advent of digital calculators, this was the fastest game in town for a trained operator. If you had a computer room full of computers using these, um you could again successfully successfully perform all of your like calculate ledgers, do all all this normal stuff that you would need to do like in an accounting firm.

So, wow, I got to adjust that. So, how would you multiply in this thing? Well, multip all multiplication is is repeated addition, right? So, I'm not going to tell you how I do it, how I'm going to do it yet. I'm just going to show how on a comtoter you multiply 12 x 12. In case you didn't catch that, let me do that one again. So, what'

I do? Um, multiply 12 twice or added 12 twice, shift it, and added 12 again. And there we go. 144. All right. So, let's skip ahead a little bit to the 1880 1888 to Burroughs company. Oh, boy. Look at this beauty. Um, so this is this is made of cast iron. Yes, it is very heavy. Is it has glass? Um I did not bring this one to

scale. Uh this is the Cadillac of business calculators. Okay. Like if you've ever watched It's a Wonderful Life when they go into Bailey um building and loan. If you're like me, you pause it and well, you don't have to pause it because you already know it's there. But there's a scene where they pan across the bank, right? The building and loan. And one of these is sitting

right there front and center because this is what banks would use to perform calc to do math. And like the thing about one of these is it has, you can kind of see it in the back. It has basically like a full typewriter um setup in the back. They had different ones, but you could put a full sheet of paper, a full ledger in the back of

these and roll it up and it you would generate records, right? The whole point of this was to put mortgage applications and other things on this. And some fancier models of this even have like buttons for months of the year um so that you could print those out as well. So this is a class one adding listing machine. Um what do I mean by adding listing machine?

Well, it can do two things. It can add and it can list what it added which means it printed it on a piece of paper. All right. An adding machine that's not an adding listing machine just adds. All right. But an adder listister has that piece of paper. This one um has green felt on it. You can kind of see it's still green felt in a black

case. This set the style of uh calculators for decades afterwards. In fact, you can kind of see next to it, you see this other calculator that's black and has green. No one this model, including Burrows, actually use uh green felt because it's kind of expensive. They just paint it green. Why' they paint it green? Because this model had green felt and that was fancy. Um and so

all the other models, including clones, the one next to it, it's not even a burrows. is just a clone from like decades later, but basically all the calculators were either solid black or they were black with a green field in the front. So, the other thing is there what's so cool about this one is there's glass in the front and the sides. Um, and if you look,

you can see there's like you can see the mechanisms and the working. If up until recently when I was preparing this talk, I thought that was just because it looks so sweet. Um, because it does. That's not why. Apparently, this was to counter the skepticism of customers who would come in and see someone using this new fangled box and entering numbers and doing calculations and they wanted

to show no there's mechanisms underneath that are performing correct calculations. Um, so they actually expose the mechanism the inner workings for that reason. Burroughs was kind of the IBM or Microsoft of its day depending on what eras you're talking about for those companies. Um and what what I mean by that is they had gigantic product lines of all kinds of different calculators um including um like this

but what they would do was they would acquires because this this was a patent heavy space. So someone would come up with a new improved design or some sort of thing they would patent it and because it's a mechanical patent no one else someone else would have to come up with another mechanical way to do it that didn't violate the patent and so we come up with

a better one. Burls would just buy the company rebrand their products as so um I don't have this there's this really cool ad I don't have and I couldn't find a good example of it that shows like 10 different burrows calculator of all the different they have all the sizes like this big Cadillac they go all the way down these little desktop models um so how do

you this this uses what I would call the standard adding listing machine UI all right so press number add once you do that you pull the lever that creates the mechanical force you need to add to the register and to print onto the listister um in the back. Uh if you want to do multiplication, there's a little repeat button that you hit because normally when you pull

the lever, it clears whatever keyboard. If you hit the re the repeat button, you can just crank crank and every time it keeps adding that same number without clearing um clearing the numbers. If you press subtotal and pull the lever, um it'll print the current total without the registers. So, if you just kind of want to see where you're at, um you can also just look down

numbers down there. But if you just want to print it out, if you actually want to clear the registers, the only way to do that is to press total, pull the lever, um and it'll print the total and then clear the registers. You can off start from scratch. Um, you see these little red buttons on the top? That's if you want to clear Want that number to

be pressed anymore. You made a mistake. You can hit the red button. It'll unpress that button. This doesn't subtract. This doesn't divide. It adds. And because it adds, it can multiply. But that's it. All right. Because it can add and multiply. Let's uh 12. There we go. All right. So, yeah. Um, adding 100 added multiplying 12 by 12 on a Burroughs class one from 1917. All right.

Sunstrand. Let's skip ahead to the to the 20th century. Now we're getting pretty advanced. Um this is a Sunstrand adding machine. This is the first adding machine that had like the standard what we would consider the standard 10 key UI that we're all familiar with. Um it had an internal register that would track what buttons you pressed in what order and then you would uh press the

numbers in sequence. You and uh if you wanted to multiply you'd press the repeat button just like the other one. No subtraction division in early models. Later models did add subtraction and division. Well, subtraction and then via that some level of division. Okay. So, I don't have one of these. Um, it's very sad. Uh, I really should uh because it's one of the few examples of like

this UI. I don't really have one of this. So, that's just a note for myself later uh that I need to get one. So, I don't have an example of multiplying 12 times 12 unfortunately. So, let's skip ahead to 1912. the Monroe. So, this beast is pretty big. It's almost It's almost as heavy as the Burrows, but it's not because it's not made of cast iron. Um,

uh, so the Monroe the Monroe Model F, which is what you're looking at here, was an attempt to appro improve the UI of the Odor Pinwell calculators. You remember the firsts I showed you that had like the sliders and stuff? They realized, you know, those sliders are kind of a pain. It'd be really cool if we like the mechanism, but we we want to make a better

UI. So, what we're going to do is copy the UI mostly from some of these other adding machines that have this full keyboard. So, we're going to take a full keyboard and combine it with the mechanis the pin wheel mechanism and we'll have a new calculator. So, yeah, it's very similar to pin wheel mechanically, uh, but it has a full keyboard instead of the sliders. So this

is not as fast as but it actually really took off because it's faster to learn. So it's kind of like the nano of uh adding machines, you know. So so you did you still needed some training. Um it wasn't as fast, but you could learn it way faster. This can actually perform So you press numbers on the keyboard to select them. You rotate the crank clockwise if

you want to add, counterclockwise to subtract. Um, there's a little knob. If you want to shift left and right when you're multiplying and dividing, you rotate that knob and it kind of lifts the back register up and shifts it over one way or the other. And let's do some multiplication. This is my model F. This is from around 1915. Now you notice in this case I didn't

have to res the numbers. The reason set all I had to do was shift the register to the right. Um that's an efficiency gain right? So let's play that again so you can compare that to some of the other ones. And if you were careful, you may have heard the ding um when I rotated that. That's because there's that made enough vibration to ding the bell. This

has a bell, too. So when you're doing division, it did the same thing where you subract, subtract, underflow, bell, ding, add one, and shift. And because this I don't want this to be a two-hour talk. I didn't do any long division on any of these. Oh, actually with one exception. U not this one. uh the facet 1932. So this the pin wheel calculators are very popular. UI

was weird. What you saw before was not the only attempt to improve the UI to be very intuitive or more intuitive than the sliders. Um Facet also uh had an attempt looks like this. Um you're like what? Yeah. So this is attempt from a Swedish company to in improve the pinwill UI. The pinwill calculators were very popular in that area but they wanted to improve the UI.

they took a different approach from Monroe um in that they did do 10 keys and you did hit them in sequence a lot like the Sunstrand but unlike the Sunstrand I mean look at that order uh they made up kind of their own order uh for this UI so you just sort of learn where they are um and press those buttons and hope that you didn't mess

up all right uh there's some display reg display register in front so when you hit the numbers in sequence it would show in the front what number you just entered so you can see what it was you can always clear those Um, you'd rotate the clank the crank clockwise to add, counterclockwise to sub. There's a little shift button in the front. You hit that button and since

instead of like turning that crank that I did before to flip things over, hit the button, it shifts left to shifts right depending on which button you press, which is pretty handy. Um, what's what's really nice about this one is there's a one-handed clearing of all the registers. So, on some of these you have to turn a bunch of cranks. This one's designed for one-handed operation. So

those two registers you kind of with one hand can grab both of these levers and squeeze them and it clears both of those registers which pretty um I don't have a original facet. This is a clone. This is actually a white label from RC from the late 1940s. They they were an importer of facets um into the US and so they would rebrand their RC Allen but

obviously same mechanism and everything else. So let's do some And if you notice here, what's nice about some of these calculators is they you can see the full operation here because many of these calculators not only show you the number that you're you're adding if you're fortunate or if you have a like this one shows you what you're multiplying by. It also keeps track of how many

times you turn the crank and when you shift it moves that over and increments that so you can actually see what you're multiplying by. Um, and then you obviously have the answer as well. March silent speed 1934. Look at this buddy. Yeah, this is a this is a pretty pretty crazy one. Um, so there's a lot of electromechanical calculators. So it this these are all there all

these calculators are mechanical but but there are a lot of these calculators also created electromechanical models and what I mean by that was some of the first models basically if there's like a lever or something that they just needed physical force to cause a thing to happen they would replace it with a motor. You'll see a lot of Burrow's ads that same exact model but they just

have like a electric motor strap to it and you'd plug it in. All right. And they may have added maybe an extra button to engage that motor but that's pretty much it. However, eventually some calculators were designed completely around the electromechanical approach, including this one. So, this is the merchant silent speed. Um, I put that in quotes, not just because it's a title, but silent is relative.

Um, but it's a this is an incredibly sophisticated example as someone who's taken some of taken it apart. Um, there's like this crazy set of nested planetary gears in the register. So when you see it operate in a minute, you're going to notice that some of the registers kind of like go halfway through sometime before they settle. And that's because there's these crazy nested planetary gears that

cause it. That's how it handled um carrying the one as it were. Uh uh this allows because it's electromechanical, this allows direct multiplication and division using buttons. It actually has a multiplication button. It has a division button on. You don't have to just do repeated addition, which is pretty nice. Uh this is also used in the Manhattan project along with like slide rules and other stuff. But

this was actually a model that was part of that project. Not the one I own but this type of calculator. Um these are vulnerable to divide by zero errors actually. Um so if you were to enter a number and then just hit divide um you're dividing by zero. You haven't entered anything to divi a divisor. Um so it just starts cranking and it starts cranking and it

doesn't know what it just keeps going. It keeps going and many labs, in fact, you start reading the history of these, many um labs at universities that would have these would have big problems with students burning them out because they would divide by zero and then like not know what to do and just sort of leave or whatever um and burn the motor out. These actually have

an emergency oh no, I did it button that you can press. It's a it's a red button and you hit the red button and it actually just stops all operations. It's just sort of like nope and then you're done. All right. So, whoa. Let me scoot ahead here. Okay. So, this one I'm not just going to multiply 12* 12. I'm actually going to calculate pi. So, I'm

going to use a sort of a traditional approach of approximating pi, which is you divide 355 by 113. So, this is how you would approximate pi using a margin silent speed. Shift all the way. There you go. With a remainder of 23. Uh, so let's try that again. Just in case you didn't catch the first time, shift all the way over to 355. Enter 113. Hit the

divide all the calculators weren't necessarily desktop calculators. Portable calculators are also existed. I'm not going to dwell as much on this as some of the other ones. Um these portable calculators, but for example, there's the Adiator. I have I have an example here. 1923 was when they were invented. Um this was designed to go in your shirt pocket. Uh, this was known as like a pin wheel

or or a shepherd's hook um design, shepherd's hook mechanism. And what I mean by that is you'll notice kind of on the top, and you can come afterwards if you want to play with one, uh, there's like a hook shape. So, when to add with one of these, you take a stylus out and you put it in the number you want to add and drag down. Unless

when you put the the stylus in the number, it's red. Or when you drag it all the way down, you get like this warning red thing on the top that says you overflowed like or and then you're like, "Oh, whoops." So, I have to drag it and carry literally carry physically carry the one over to the over the shepherd's hook and it will carry and it'll carry

the one to the next register. All right. Um, this one in particular has two sides. So, if you want to add, you put the numbers in this side and when you're ready to subtract, you flip it and then you do the same thing on the other side. All right. These are great with Scrabble. Actually, this is my favorite way to tally up Scrabble scores. There's also the

adometer. Uh there's a series of different types of models of adometers. This was uh invented in 1927. Uh these were more like a rotary dial and you would stick a stylus in and start rotating and it says it the instructions actually on the top which is pretty nice UI to tell you how to use it. You would rotate um counterclockwise to uh add and you would turn

clockwise Did I get it wrong? Yeah. Um you'll turn to left. Well, that's not very helpful, is it? Uh to subtract it. You Yeah, I got it wrong. So, you drag it all the way down. Um and so anyway, but what you notice is there's two sets of numbers. So, the bigger numbers is what you where you put the stylus to add and the smaller number is

where you'd put it if you were subtracting because of the internal mechanism. That's just how it had to work to physically work. You had to you had to know how to do that. If you squint on this one, you'll notice something strange. Um the first sets of digits are 0 through nine. Um then it goes to 11. Uh which is pretty sweet. Uh uh didn't stop at

10, didn't stop at nine. It is it is one louder. Yeah. Um why did it why does it go to 11? And if you if you go to the go to the right of the 11, you'll notice it has fractions. All right. And this is this is pretty common. A lot of these these handheld ones would help you with fractions and they'd have a common fractions. You

could add those. This was designed for carpenters who were meas doing doing calc adding measurements like physical measurements of things they were working on. Um that particular the one that goes to 11 is measures inches. So the idea is you're adding inches and when you carry the one in inches you have a foot. So the one next to the 11 is feet measurements. And there's all kinds

of different ones that for like the the British currency before standardization and all of that stuff. There's all kinds of stuff like that too which we won't get into. Oh, the Curta. I love the Curta. Uh 1938 probably this engineer portable calculator. Many would say the finest port calculator period. I'd probably agree, but I got to be careful. A bunch of geeks out here. You say something's

superlative. Some I have a counter example. So this is those that original odner you that pin wheel mechanism shrunken down and condensed into a portable pocket-siz container. Some people call this like a math grenade. Uh it's about that size. This can perform the four main types of arithmetic. You can and expand on that to do all the other things. Uh you slide the sliders kind of like

the pin wheel calculator, but you slide these sliders to set the digits. You turn the crank clockwise or counterclockwise depending on what operation you're doing. You lift the center and twist to shift left or Um I don't have one of these. Uh these are really expensive. Uh if you don't know, like like a lot of geeks really like this model. I mean this I get it. It's

really sweet. Um they're like they start at like $1,000. Uh whereas most of these other things most of these other calculators in my collection, you can probably find them for like somewhere between 20 and 50. Some really nice examples maybe a hundred bucks. Um sometimes more. But yeah, these start at like a thousand um and and go up from there for sure. >> I the question was

how much did they cost when they originally sold? I don't know. Um there's also a cool interesting story about this which I won't get into because it's this is a UI talk. Uh now let's skip to the 60s and now we're entering the digital age. So goodbye clicketity clack. Um so digital calculators started appearing in the 1960s. Uh these were much quieter than mechanical calculators to start

with. All right. Um they're faster than electromechanical calculators were at performing calculations. Way fewer moving parts and much less maintenance. Something that you may not realize is that along with these adding machines that you and calculators you would buy as a business, you get a maintenance contract and you would have people that would come to those offices just like a typewriter repair person who would come to

the offices and routinely need to perform maintenance, oil things, repair things. I mean, I have a different model that I brought to scale to demo. And I was showing Bill last night and started cranking and one and one of the registers now needs to be repaired. like I was trying to reset and it got there's basically an internal clutch that got stuck I'm pretty sure because has

it has old oil and grease and over time that gets kind of uh sticky and so nine times out of 10 if you get into this hobby and you want to repair these things because they don't work all you need to do is sort of take them apart a little bit clean them and oil them and then magically they start working again but you still have to

do that digital calculators nope they there's no oiling of a digital calculator the other thing that's interesting is the UI UI is now for the first time decoupled from the mechanism. The person designing the UI for these calculators didn't have to think about well how how do I engage this mechanism within the calculator to then do all like turn gears and do all this stuff? How am

I going to power it? All they really had to think about was they actually started thinking more about the user. Um how is the user going to how should they input numbers into this thing? And if you notice, they pretty much settled on the 10 key UI that we still use today. Um, they felt at the time that that's probably the most intuitive or at least the

easiest to use of the UI. So, what does this all mean? Uh, lessons for starters, I'm not a UI expert. All right? I don't pretend to be. Uh, but what I've gathered from my time talking to people that are interested in this and being a user myself is that empathy seems to be really important when you're building UI. You actually need to think about things from the

other person, the user's perspective instead of your own. Um, the thing is it's natural to develop to to treat your UI as just an extension of the code underneath, but uh you've seen some of the consequences if you only expose internal mechanisms to the user and you're not really thinking about how they're going to use it. So some suggestions I would if you're building something that has

a CLI I would avoid a CLI that basically just mirrors the internal functions and data structures that you know and just exposes them directly to the user. All right, think about how they're um going to use it. So I'm kind of looking at UTC and IP and there's plenty of other examples but I felt safe calling those out um as some examples. I would say take take

advantage of muscle memory and conventions when it makes sense. There are certain things some people just sort of like I'm just going to experiment with random UI. I think it looks nice or whatever. Um that completely goes against conventions. Uh sometimes those that muscle memory those conventions are very beneficial. Someone comes through your software completely with not not knowing how to use it. It's nice if things

are maybe where they might expect them to be. So for example, why are like all the core Pac-Man flags uppercase? Uh why uppercase S for like the main way that you would install software? Uh why uppercase R to remove packages? Like and like I don't know I I'm not a big Arch user. I just like every time I run across the documentation because I'm on Arch's website

all the time because their docs are sweet. Um as everyone knows they're like great great docs. But I watch Pac-Man stuff. I'm like I don't get it. Like why is it uppercase to begin with? Like you know. Anyway, small rant. This said be willing to leave conventions behind. Not everything that we do is always the right. The way that we've always done it isn't necessarily right. Um,

you also need to balance this advice with the willingness to maybe change the UI if you're like maybe the way we've always been doing it kind of sucks. All right. Um, main question is how far can a user get with your software with without a manual with no training? If you just handed someone their software, could you could they use it? Um that's basically I would say

the foundational question for your UI in my opinion. So for example um how many Linux installers still basically expect you to have deep knowledge of the DRO uh to you to install it? Does that make sense for a new user coming into the DRO who doesn't actually like I'd like to try this DRO out. Okay. Well, first let me develop my deep dro knowledge so I can

answer the questions to install the DRO so that I can then develop the deep deep dro knowledge. Right? doesn't make sense. But here's another thing is what if sort of like the digital calculators, what if you could actually completely decouple how how your software works from how it's used. One of the reasons that digital calculators settled on that 10 key UI at the time was because they

no longer had physical constraints. They were no longer bound by well the mechanism is this big and it has to fit in this shape and we have to be able to have a lever when you press a thing to trigger this other thing. They didn't have to worry about that. that they just know I have buttons and it and it it sets things electronically. So they could

they didn't they could think about the UI completely independently from how it operated under the hood. And I guess I would finally leave leave you with this. Even the current calculator UI, the thing I opened up with u isn't intuit based on technical limitations. The reason that when you open a calculator app on your phone and it has that same thing is because it's how they always

did it at least for a couple of decades. Um, and the reason they always did it that way was that was the best way to technically input those numbers when digital calculators appeared only, right? That's not necessarily the best way to input numbers on a smartphone or something that doesn't have those physical limitations. For example, um you give a kid a calculator for the first time, it's

not necessarily intuitive how to perform arithmetic on it, right? Um many pe many kids, if you were to have them do arithmetic, they probably open up a search bar somewhere, type in the functions, and get an answer. Or better, they might even just talk to a voice assistant, give it the arithmetic, and have the answer come back. Right? like that seems more like a more basically what

I'm trying to say is um decouple yourself from the technical limitations and think about what the user interface could look like if those technical limitations went away even for Uh thank you very much and open for All right. Any any question? Oh, yeah. >> Do you know why the first digital calculators decided to use the 10 key mechanism? Was there any specific design constraint or reason that

they decided to do it? I don't know of any particular like I don't have access to any sort of discussions or determining where I saw them determine that. I think guess this is just a guess is that they looked at the available UIs that they could draw from everything from the comptoter which is still being made um all the way to all the other things and I

think they looked at the sun strand and they said was from like zero one through nine which is we don't we want to make and I think that's why they settled on it more than anything. Any other questions? >> Yeah. is uh what do you think about HP and reverse polish notation? Was that an honest attempt to improve the UI or were they going in another direction?

>> I think I think reverse polish notation was an attempt to provide something that was helpful for a power user not that was intuitive for anyone who would learn math traditionally. >> Hey, great great talk. So um you know as you were going through your slides uh the the the movie um imitation game came to my mind where where that with the enigma being kind of a

mechanism for doing similar stuff mechanically as well. I'm just wondering if you ever took a deep dive to understand how that system works to implement encryption. >> Yeah I never Oh the >> the enigma is another missing piece of my collection. Of course now you I mean they're millions of dollars right? Uh, I didn't do a deep dive into how the Enigma worked. I've looked at it

a little bit. There's other really cool uh, cryptographic mechanical calculators that are out that are out there from you can still for only a few thousand or maybe a few hundred dollar. Sometimes you can find some of those that were like lesser used. They weren't like they weren't actually the enigma that's like some other country use it post World War II or something. Um, yeah. I mean

I Yes. So, I haven't looked into it too deeply. Um, I would love to have one. And I would love even have a reproduction of one. Uh but they're really complicated to make apparently and like the machining cost of like rebuilding some of one of those things apparently is the reason that you don't see a lot of them uh anymore. >> Yeah, it's open knowledge and how

it's made, but recreating them requires like the mechanical precision um to create all the individual parts by hand. We would massproduce it, but we'd have to It would be incredible cost to create the dyes and molds and all this stuff because it's kind of Yeah. to machine all the parts. >> I'm just wondering about where did you find all these antique calculators? Because I'm interested in antiques

myself, but I kind of would like to find some myself. >> Yeah. So, um, where I found these is a blend of basically three places. So, antique stores are a great place to find these. They will show up. Um, I haven't checked the antique mall down the street here uh since I've been here. But yeah, so often in the past I've seen one or two there. Um,

in addition to sometimes thrift stores will still have them. Um, in particular, handheld units like these kinds of things you'll see turn up in thrift stores quite a bit or flea markets or things like that still. Um, eBay is another great choice. And I actually have part of my collection I got from Craigslist. Actually, a a major about four pieces of my collection. Some of my prime

pieces I got from a collector who passed away and his wife had a garage full of I kid you not like 300 typewriters. Um, and probably a hundred adding machines and calculators of different vintages. Probably like four or five comptoometers. That burrows I showed you, probably three or four of those. And she's just like opening it up for, you know, like come and buy these things. And

she had someone appraise them and stuff. So I went through and bought like three or four from this sort of like a estate sale >> before before computers became popular. I know that there was um something called a real programmable paper computer from Instruill. Have you ever heard of those before? >> I have not heard of that. No. So yeah, I have haven't heard of the programmable

paper computer, but yeah, thanks for for mentioning that. Yeah. >> Anybody else? Well, thank you very much for wonderful. >> Thank you everybody. And of course, talk in the hall too.